Skip to main content

Full text of "The American journal of science"

See other formats


\ ^ % % <$> mgm % # m 

£ft5£p ^ ^ 



A ./, 





^ A 



^ ^^rg^ 


^ o\ -- - :-- - 

Jfe - _=r-€-_ - _ ... J? ^%^ *\ J?^ jj 

a n ^^ rP ^> X° ^~~^ V 


cf % f 

' /j ^ ^ if f ^ / 1 


■y 3 o 6. J^> 




Editor: EDWARD S. DANA. 



op Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 

of Washington. 




^ 6 -2. C+ 2- 



* J 



Number 19. 


Art. I. — The Melting of Potash Feldspar; by G. W. Moeey 

and N. L. Boavex, 1 

Art. II. — Triasssic Reptilian Order Thecodontia, by F. 

vox Huene, 22 

Art. III. — A Discussion of Triple Salts; by H. L. Wells, 27 
Art. IV. — Horned Eocene Ungulates; by E. L. Troxell, . . 31 
Art. V. — The Genus Hyrachyus and its Subgroups; by E. 

L. Troxell (with Plate I), 38 

Art. VI. — A New Occurrence of Ilsemannite; by C. W. Cook, 50 
Art. VII. — On the Zonal Division and Correlation of the 
Silurian of Bohemia ; by J. Perner, with the collaboration 
of O. Kodtm, 53 


Chemistry and Physics. — A New Process for the Industrial Production of Barium 
Hydroxide for Treatment of Molasses in Sugar Kenning, Deguide and Bode: 
Advanced Course of Instruction in Chemical Principles, A. A. Notes and 
M. S. Sherrill, 73. — Determination of Sulphur in Iron and Steel, H. B. 
Pulsifer: Organic Chemistry, V. v. Richter, 74. — Friction and Lubri- 
cation. Hardy and Doubleday, 75. — Power Alcohol, G. W. Monier- 
Williams, 76. — The Journal of Scientific Instruments: La Theorie Ein- 
steinienne de la Gravitation, 77. 

Geology and Mineralogy. — Gravity Anomalies and their Geological Interpre- 
tation, 78. — Publications of the United States Geological Survey, G. O. 
Smith, 79. — Die Eruptivgesteine des Kristianiagebietes IV; Das Fengebiet 
in Telemark. Norwegen, W. C. Brogger, 80.— Mineral Resources of the 
Philippine Islands for 1919 and 1920. 82. -A List of new Crystal Forms 
of Minerals: Handbook and Descriptive Catalogue of Gems and Precious 
Stones in the U. S. National Museum, G. P. Merrill, etc.: Virginia Geo- 
logical Survey, T. L. Watson: The Topographic and Geological Survey 
of Pennsylvania, G. A. Ashley: Geology of Drumheller Coal Field, Alberta, 
J. A. Allan, 83. — Potash in a new area of Texas, 84. 

Natural History. — Arctic Alcyonaria and Actinaria, A. E. Verrill: Genetics, 
An Introduction to the Study of Heredity, H. E. Walter, 84. — A Natur- 
alist in the Great Lakes Region, E. R. Downing: La Constitution des 
plantes vasculaires revelee par leur Ontogenie, G. Chatjveaud, 85. — The 
Vegetation of New Zealand, L. Cocayxe: Les Mouvements des Vegetaux, 
R. Dutrochet: Die Pflanzenwelt Afrikas, insbesondere seiner tropischen 
Gebiete, A. Engler, 86 — Precis de Physiologie Vegetale, L. Maquenne: 
The North American Slime-moulds, T. H. Macbride, 87. — Soil Conditions 
and Plant Growth, E. J. Russell: A Handbook of the British Lichens, 
A. L. Smith. 88. 

Miscellaneous Scientific Intelligence. — The Outline of Science, J. A. Thompson, 
88. — Publications of the Smithsonian Institution, G. D. Walcott, 89. — 
Banking, Principles and Practice, R. B. Westerfield: Civic Science in 
the Home. G. W. Hunter and W. G. Whitman, 90. — Memoirs of the 
Queensland Museum: United States Life Tables. J. W. Glover, 91. — Public 
Opinion, W. Lippmann, 92. — Publications of the Carnegie Foundation for 
the Advancement of Teaching, 93. — American Association for the Advance- 
ment of Science: Observatory Publications, 94. 

Obituary.— G. Simonds: A. Bacot: L. A. Ranvier: H. M. Howe, 94. 


Number 20. 


Art. VIII. — Colloids in Geologic Problems ; by G. D. Hubbard, 95 

Art. IX. — Primitive Pecora in the Yale Museum; by R. S. 

Lujll, , Ill 

Art. X. — A Critical Phase in the History of Ammonites; by 

C. Diener, 120 

Art. XI. — Saccoglottis, Recent and Fossil; by E. W. Berry, 127 

Art. XII. — A Crossotheca from the Rhode Island Carbon- 
iferous ; by E. M. Round, 131 

Art. XIII. — AFossil Dogwood Flower; by F. H. Knowlton, 136 

Art. XIV. — Intrusive Rocks of the Portsmouth Basin, Maine 

and New Hampshire; by A. Wandke, 139 

Art. XV. — Babingtonite from the Contact Metamorphic 
Deposits of the Yakuki Mine, Province Iwaki, Japan; 
by M. Watanabe, 159 

Art. XVI. — ATillite-like Conglomerate in the "Eocambrian" 

Sparagmite of Southern Norway; by O. Holtedahl, . . . 165 

Obituary.— A. G. Mayor, 173. 


Number 21. 

Art. XVII. — The Determination of the Space Group of a 

Cubic-Crystal; by R. W. G. Wtckoff, 175 

Art. XVIII. — The Symmetry and Crystal Structure of Zinc 
Bromate Hexahydrate, Zn(Br0 3 ) .6H„0; by R. TV G. 
Wtckoff, : .,. 188 

Art. XIX. — On the Symmetry and Crystal Structure of 
Sodium Hydrogen Acetate, XaH(C o H 3 2 ) 2 ; by R. W. G. 
Wtckoff, " 193 

Art. XX.— Cone-in-Cone; by W. A. Tarr, 199 

Art. XXI. — Xotes on the Flora of the Payette Formation; 

by R. W. Chaxey, 214 

Art. XXII. — Xotes on the Structure of the Triassic Rocks 

in Southern Connecticut ; by C. R. Loxgyvell, 223 

Art. XXIII. — Amphisymmetric Crystals; by E. T. Wherry, 237 

Art. XXIV.— A New Trilobite Appendage; by T. H. Clark, 215 

Art. XXV. — Cyprine and Associated Minerals from the Zinc 

Mine at Franklin, X. J. ; by J. V. LEWisandL. H. Bauer, 249 


Miscellaneous Scientific Intelligence. — Stratigraphy of Northwest Greenland, L. 
Koch, 251. — Revue cleGeologie et des Sciences connexes: First Pan-Pacific 
Commercial Conference: First Congress of Industrial Chemistry, 252. 

Obituary. — A. G. Bell, 252. 


Number 22. 


Art. XXVI. — Jones's Criticism of Chamberlin's Ground- 
work for the Study of Megadiastrophism; by T. C. 
Chamberlin, 253 

Art. XXVII. — Relation of Sea Water to Ground Water 

along Coasts ; by J. S. Brown, 274 

Art. XXVIII.— A Petrologic Study of the Cape Neddick 

Gabbro; by A.Wandke, . . 295 

Art. XXIX. — Fossils of the Olympic Peninsula; by W. H. 

Dall, 305 

Art. XXX. — A Mid-Devonian Callixylon ; by C. J. Hy- 

LANDER, 315 


Chemistry and Physics. — A Kevision of the Atomic Weight of Beryllium: The 
Analysis of Beryllium Chloride, O. Honigschmid and L. Birlenbach: 
Experimental Attempts to Decompose Tungsten at High Temperatures, 
G. L. Wendt and C. B. Iron, 322. — A Micro-Method for the Determina- 
tion of Molecular Weights in a Melting-Point Apparatus, K. Rast: 
Women in Chemistry: A Study of Professional Opportunities, The Bureau 
op Vocational Information, 323. — Spectrum of Aurora, 324. — Suspended 
Impurity in the Air, 325.— The Principles of Geometry, H. F. Baker, 326. 

Miscellaneous Scientific Intelligence. — Smell, Taste, and allied Senses in the 
Vertebrates, G. H. Parker: Science and Human Affairs from the View- 
point of Biology, W. C. Curtis, 327. — The Biology of the Sea-Shore, 
F. W. Flattely and C. L. Walton, 328. — New Meteoric Iron from Ken- 
tucky, G. P. Merrill, 329. 

Obituary. -E. D. Salisbury: G. H. Cox, 329. 


Number 23. 

Art. XXXI. — The Silicates of Strontium and Barium: by 

P. Eskola, 331 

Art. XXXII. — Sedimentation in Lake Louise, Alberta, 

Canada ; by W. A. Johnston, 376 

Art. XXXIII. — Imbricated Structure in River-gravels ; by 

W. A. Johnston, 387 

Art. XXXIV.' — Zircon as Criterion of Igneous or Sedimen- 
tary Metamorphics; by P. Armstrong, 391 

Art. XXXY. — The Minnesota Devonian and its Relationship 
to the General Devonian Problem of North America; by 
C. R. Stauffer, 396 


Chemistry and Physics. — A New Method of Separating Arsenic from All Other 
Metals, L. Moser and J. Ehrlich: A New Volumetric Method as Applied 
to Certain Problems in Inorganic Chemistry, P. Dutoit and E. Grofet, 
413. — Theories of Organic Chemistry, F. Henrich: The Chemistry of 
Combustion, J. N Friend. 414 — Petroleum, Where and How to Find it, 
A. Blum: The Heavier Constituents of the Atmosphere, J. J. Thomson: 
The Corrosion of Iron and Steel, 4 15. — The Mathematical Theory of Prob- 
abilities, A. Fisher, 417. 

Geology. — The Paleontology of the Zorritos Formation of the North Feruvian 
Oil Field, E. M. Spieker: The Recession of the last Ice Sheet in New 
England, E. Antevs, 417. — A Section of the Paleozoic Formations of the 
Grand Canyon at the Bass Trail. L. F. Noble: Essentials for the Micro- 
scopical Determination of Minerals and Rocks in Thin Sections, A. 
Johannsen: The Rocks of Mount Kverest, 419. — A Newly Found Tennes- 
see Meteoric Iron, G. P. Merrill: Minor Faulting in the Cayuga Lake 
Region, E. T. Long, Errata, 420. 

Miscellaneous Scientific Intelligence. — Foundations of Biology, L. L. Woodruff: 
The Study of Living Things: A Course in Biology for Secondary Schools. 
W. H. D. Meier, 421. — Field Museum of Natural History, Annual Re- 
port for 1921: National Academy of Sciences, 422. 

Obituary — A. Smith: F. T. Trouton: D. Sharp: W. Kellner: A. L. Kimball: 
A. A. Sturley, 423. 


Numlber 24. 

Art. XXXYI. — John Day Felidse in the Marsh Collection; 

by George F. Eaton, 425 

Art. XXXVII.— The Antimony Mines of Shin Chow, China; 

by G. D. Hubbard, , 453 

Art. XXXVIII. — A Tribolite retaining Color-Markings; by 

P. E. Raymond, 461 

Art. XXXIX. — On the Occurrence of Richthofenia in Japan ; 

by I. Hayasaka, . 465 

Art. XL. — On the Crystal Structure of Ammonium Chloride; 

by R. W. G. Wyckoff, 469 

Art. XLI. — The Alleged Variable Composition of Triple 

Chlorides Containing Silver and Gold; by H. L. Wells, 476 

Art. XLII. — The Structural and Stratigraphic Relations of 
the Great Triassic Fault of Southern Connecticut; by 
W. L. Russell, 483 


Chemistry and Physics. — The Presence of Cobalt and Nickel in Vegetables, 
Bertrand and Mokragnatz: Standard Methods of Chemical Analysis, 
W. W. Scott, 498. — Outlines of Theoretical Chemistry, F. H. Getman: 
The Formation of Colloids, T. Svedberg: Physikalische Chemie der Zelle 
und der Gewebe, E. Hober, 499. — Spectral Determination of Temperature, 
I. G. Priest: Vector Calculus, J. B. Shaw, 500. — The Origin of Spectra, 
P. D. Foote and F. L. Mohler, 502. 

Geology and Mineralogy. — Pottsville Fauna of Ohio, H. Morningstar, 502. — 
Earth and Star, an Hypothesis of Weather and Sunspots, E. Huntington 
and S. S. Visher: Seventeenth Report of the Director of the New York 
State Museum and Science Department, 503.— Sveriges Olenidskiffer, A. 
H. Westergard: The Geology of the Broken Hill District, E. G. Andrews, 
504. — Eclogites of Norway, P. Eskola, 505. — New Deposits of Radium in 
Afrika: United States Geological Survey, P. S. Smith, 506. 

Miscellaneous Scientific intelligence. — National Academy of Science, 506. — 
Nobel Prizes for 1921 and 1922, 507. 

Obituary, — J. Waterhouse: W. H. Wesley: C. M. Smith: L. A. Tchcjgaiev: 
R. W. Wilson: C. W. Waggoner, 507. 

Index, 508. 

VOL. IV JULY, 1922 

Established by BENJAMIN SILLIMAN in 1818. 



Editor: EDWARD S. DANA. 



of Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 

of Washington. 



No. 19 -JULY, 1922. 



19 22. 


Published monthly. Six dollars per year, in advance. $6.40 to countries in the 
Postal Union ; $6.25 to Canada. Single numbers 50 cents ; No. 271, one dollar. 

Entered as second-class matter at the Post Office at New Haven, Conn., under the Act 
of March 3, 1879. 


Huntington & Williams' 


By Ellsworth Huntington 

Research Associate in Geography, Yale University, 

and Frank E. Williams 

Professor of Geography and Industry, Wharton School of Business, Uni- 
versity of Pennsylvania. With the co-operation of 

Robert M. Brown, Professor of Geography, Rhode Island College of 
Education, and 

Chase, Teacher of Geography, Mount Vernon (N. Y.) 

Miss Lenox E. 

High School. 

This book was written by an exceptionally strong combination of authors 
to meet the urgent and long recognized demand for a really practical Geogra- 
phy text around which a ''Business" course in this subject could be built. It 
has been prepared especially for use in commerce departments in high schools 
and Freshman courses in colleges, and will prove an ideal text foi such work. 
It is scientifically sound throughout, is written in an intensely interesting way, 
and answers the questions the business man would ask, in the language the stu- 
dent understands. 

August 1st, in plenty of time for the fall classes. It will contain about 470 
pages, 6 by 9, well illustrated, and will be bound in cloth. The price is not 
yet known. 

And A New Edition — The Second — of 


This epoch-making book, which marked a new era in the teaching of 
Geography, has been carefully revised and now appears in materially improved 
form. It is used in about 150 schools, colleges and normal schools. 

* 430 pages. 6 by 9. Illustrated with photographs and maps. Cloth, $3.50 


432 Fourth Avenue 
New York 

London Montreal, Quebec 

CHAPMAN & HALL, Ltd. Renouf Publishing Company 

AJS 6.22 

T H E 



Art. I.— The Melting of Potash Feldspar; by G. W. 
Moeey and N. L. Bowex. 


While studying systems of the alkaline oxides with 
silica, alumina, and water, the one of us (Morey) carrying 
on the thermal and chemical work and the other (Bowen) 
the optical examination, we succeeded in crystallizing 
some pure artificial orthoclase in a bomb with no flux 
other than water. Crystalline orthoclase, as hitherto pre- 
pared artificially, has been contaminated somewhat by 
the flux used in preparing it, and for this reason natural 
crystals have always been used in determining the melting- 
point of orthoclase. The natural crystals never have the 
theoretical composition KAlSi 3 8 but always contain a 
considerable amount of albite and sometimes other mole- 
cules. We concluded, therefore, that the pure crystals 
we had prepared would furnish good material for the 
redetermination of the melting-point of orthoclase, a 
point of much importance in the system under investi- 

The melting-point of orthoclase, as determined on 
natural crystals, has been found to be in the neighborhood 
of 1200" C. It has been found by Day and Allen to melt 
very sluggishly. 1 We expected, therefore, that long 
exposure to a given temperature would be necessary, 
before assurance of the attainment of equilibrium might 
be had. For such long exposures we have had the inesti- 
mable advantage of the use of a furnace-temperature 
regulator designed by our colleague, Mr. H. S. Roberts. 2 
By means of this regulator we were enabled to maintain 

1 The isomorphism and thermal properties of the feldspars, Carnegie Insti- 
tution of Washington, Pub. Xo. 31, pp. 51-53. 

2 J. Wash. Acad. Sei., 11, 401-409. 1921. 

Am. Jour. Sci.— Fifth Series, Vol. IV. No. 19.— July. 1922. 

2 Morey & Bowen — Melting Potash Feldspar. 

a constant temperature (varying less than one degree) 
in our furnace day and night, for a week or more. 

Preliminary Results. 

When we heated our artificial orthoclase for a week at 
the temperature noted above (1200°) we obtained a 
product with the appearance of a glass when examined 
megascopically, and with the low refractive index (about 
1.485) and the isotropic character of orthoclase glass 
when examined under the microscope. 

However, under a high power (600 diameters) and with 
the cone of light cut down to a small angle, this "glass'' 
is found to show a sort of structure, visible only when it 
is immersed in a liquid that matches it very closely in 
refractive index. This structure can be described only 
as a sort of fine cross-lining, usually rectangular and 
giving therefore a grating effect. Day and Allen have 
suggested that. orthoclase, on melting, loses the ordered 
arrangement characteristic of the crystalline form only 
very slowly. 3 Thinking, therefore, that this observed 
structure was inherited from the crystalline material by 
the extremely viscous liquid, we held it at a somewhat 
higher temperature (1225°), where the liquid would be 
more mobile, with the expectation that the structure would 
disappear. Instead of this we found that it became more 
distinct. At a still higher temperature (1250°) it became 
apparent that the material was not a homogeneous glass 
but was made up of two phases, the one occurring as 
skeleton crystals in widely extended, branching forms of 
rectangular pattern, and the other, of somewhat lower 
refractive index, acting as a matrix for these. This 
latter was subsequently shown to be glass, but under 
crossed nicols the whole mass appears to be doubly 
refracting with a grating structure recalling that of 
microcline. At a still higher temperature (1300°), where 
the crystals are present in smaller amount, they grow 
as more discrete grains, though retaining skeletal ten- 
dencies and appearing usually as rectangular crosses. 
Finally, when formed at a temperature of 1400° or higher 
the crystals assume definite rounded to polyhedral (icosi- 
tetrahedral) forms, with indistinct patchy birefringence, 

3 Op. cit., p. 54. 

Morey & Bo wen — Melting Potash Feldspar. 3 

with a refractive index only a little higher than that of 
the glass in which they are embedded, which is about 1.49, 
with regularly arranged inclusions of glass ; in fact, with 
all the characters of leucite crystals. The series of forms 
assumed by the leucite with increasing freedom of growth 
(increasing temperature) is shown in fig. L This series 
is strikingly similar to that observed by Pirsson in the 
small leucite crystals in the groundmass of rocks from 
the Bearpaw Mountains. 4 The more primitive skeletal 
forms are not shown in Pirsson 7 s series. 

Fig. 1. 


Fig. 1. — Grains of glass containing leucite. Showing increasing perfec- 
tion of form, with increasing temperature, of leucite crystals as grown in a 
mixture of composition KAlSi 3 O s . 

The rounded leucite crystals, obtained at higher temper- 
atures, may have a diameter as great as 0.025 mm. and 
are surrounded by a rim of strained glass which is 
observed as a birefracting halo under crossed nicols. In 
fact the birefringence shown by the glass immediately 
adjacent to the crystals is decidedly greater than that of 
the crystals themselves. The birefringence of the glass 
fades out as the distance from a crystal increases, and 
when the crystals are not numerous there are areas of 
ordinary isotropic glass. When the crystals are closely 
spaced, however, all of the interstitial glass is birefrac- 
ting, which accounts for the fact that, when the crystals 
grow as ramifying skeletal forms, the whole mass is bire- 
fracting, with the cross-grating effect analogous to that 
in microcline. 

In the birefracting halo about the crystals y is radially 
disposed, which shows that the glass is subjected to radial 
tension. _ This tension is no doubt caused by the abrupt 
contraction of the leucite crystals in cooling through their 

4 L. V. Pirsson, this Journal, 2, 145-146, 1896. 
in Bock Minerals, p. 249 (2d ed., 1911). 

Figured also by Iddings 

4 Morey & Bowen — Melting Potash Feldspar. 

inversion point in the neighborhood of 700°, where they 
experience a volume change of more than 2 per cent. 5 
This may seem a small change, but it is twice as great as 
the whole change between 585° and room temperature. 

Detailed Results with Artificial and Natural Feldspars. 

The above observations showed very clearly that potash 
feldspar has no true melting-point, that the point in the 
neighborhood of 1200°, which has hitherto been regarded 
as its melting-point, is really the temperature at which it 
breaks up into liquid and leucite or, as it is commonly 
stated, it melts incongruently. It is perhaps not sur- 
prising that the existence of leucite in the material 
obtained near the decomposition-point has not been 
detected hitherto. Megascopically the "glass" has all 
the appearance of an ordinary glass ; it is perfectly trans- 
parent, with only a faint suggestion of a bluish opales- 
cence, the leucite crystallites being too minute and too 
closely matched in refraction by the medium in which 
they are embedded to cause a scattering of light. Even 
under the microscope the leucite appears, as we have seen, 
merely as an indefinite crosslining of the "glass" and 
only when the temperature was raised, to see if this 
structure would disappear, did the structure assume more 
definite form and finally become identifiable leucite 

After obtaining these results we then proceeded to 
determine the temperature of incongruent melting more 
accurately and also to fix the temperature at which the 
leucite crystals disappear and the mass is entirely liquid. 
While carrying out these determinations we studied the 
similar changes as displayed by natural potash feldspars 
in as pure a state as we could obtain them. For this 
purpose we made use of three analyzed feldspars, micro- 
cline from Mitchell County, North Carolina, sanidine from 
Laacher See and adularia from St. Gotthard, the last 
being the nearest to pure potash feldspar, though even it 

5 So far as we are aware the volume change has not been measured, 
but it may be estimated from the change of refractive index between 585° 
and 750° as observed by Einne and Kolb, Neues Jahrb. 2, 157. 1910. 

Morey £ Bowen — Melting Potash Feldspar. 5 

contains about 10 per cent albite. The composition of 
these minerals is shown in Table I. The last two analy- 
ses were made by Dr. H. E. Merwin in connection with an 
investigation he has not yet published. The results of 
his analyses, together with some of the material, he has 
kindly turned over to us. 

Table I. — Analyses of natural feldspars, 


Microcline Sanidine 

(Mitchell Comity, (Laacher See, 

North Carolina, U. S. A.) Ehineland) 

SiO, . . . 65.83 64.21 

ALO, 18.07 19.10 

Fe~0 36 

Cab". 42 .24 

Xa,0 2.30 1.76 

K,6 13.02 14.60 



(St. Gotthard, 









I. E. T. Allen analyst. Recalculated to anhydrous basis. 
Day and Allen, op. cit. p. 48. The microscope shows that a very 
little of the albite is present as perthitic stripes. 

II and III. H. E. Merwin analyst. 

The details of the experiments on these three natural 
feldspars and our artificial material are given in Table II. 
It may be noted here that all four show the breaking up 
of the feldspar into liquid and leucite. The temperature 
at which this occurs is not greatly different in the differ- 
ent feldspars, being lowered only a little by the presence 
of albite. The evidence points to about 1170° as the 
proper temperature for pure potash feldspar. There is 
moreover no appreciable difference of behavior connected 
with the difference in form of orthoclase and microcline. 

All of the feldspars show a very large temperature 
interval in which the mass consists of leucite and 
liquid. A full discussion of the change of habit of the 
leucite with increase of temperature has already been 
given for the pure artificial material. Natural crystals 
usually give more perfect leucite forms than the pure 
material when exposed to a given temperature for an 

ture °C. 

Time of 


8 days 

6 Morey & Bow en — Melting Potash Feldspar. 

Table II. — Results of heating experiments on potash feldspars. 



Little change except formation of rare 
stripes of glass corresponding with 
original stripes of albite. 
1179 8 days Abont half unchanged microcline ; other 

half glass showing cross-lining (leu- 

All changed to glass showing cross-lin- 
ing (leucite). 

All changed to glass and definite skele- 
tal forms of leucite. 

Glass and definite skeletal leucites, often 
as rectangular crosses. 

Glass and rounded to subhedral leucites 
(0.02 mm. diameter) with regularly 
arranged inclusions of glass. 

Glass and rare subhedral leucites. 

All glass. 

Sanidine and Adularia* 

Mostly unchanged. Small percentage, 
glass showing cross-lining (leucite). 

About one quarter unchanged feldspar ; 
rest glass showing cross-lining (leu- 

All changed to glass showing cross-lin- 
ing (leucite). 

All changed to glass and definite skele 
tal forms of leucite. 

Glass and minute rounded leucites. 

Glass and very rare minute leucites. 

Glass only. 

a The results obtained with these are practically identical and can be 
given together as above. 


8 days 


5 days 


3 days 


3 days 


3 days 
3 days 


8 days 


8 days 


8 days 


5 days 


3 days 
3 days 
2 days 

Morey & Boicen — Melting Potash Feldspar. 7 

Pure synthetic potash feldspar (initially crystalline). 

No change. 

About half changed to glass showing 
cross-lining (leucite.) 

All changed to glass showing cross-lin- 
ing (leucite). 

Glass and definite skeletal forms of leu- 

Glass and definite leueites, principally 
rectangular crosses. 

Glass and minute rounded leueites. 

Glass and rounded leueites (0.02 mm. 

Glass and rare leueites. 

Glass only. 

Pure synthetic KAlSi 3 O s {initially glass). 

Glass and definite skeletal forms of leu- 

Glass and definite leueites, principally 
rectangular crosses. 

Glass and minute rounded leueites. 

Glass and rounded to subhedral leueites. 

equal period. The foreign matter apparently lowers the 
viscosity of the liquid appreciably. Notably typical 
leucite crystals were obtained by holding the North. Caro- 
lina microcline at 1400° for 3 days. 

The upper limit of temperature at which leucite crystals 
are obtained varies greatly in the different feldspars. 
As is to be expected, the presence of other compounds 
lowers the temperature of complete melting. Thus the 
Carolina microcline is all liquid at about 1440°, the sani- 
dine and adularia at about 1460°, but the pure potash 
feldspar is not completely liquid until a temperature of 
about 1530 z is reached. Therefore, between the temper- 
atures 1170° and 1530°, an interval of 360°, a mass of 
the composition of pure potash feldspar, KAlSi 3 O s , con- 
sists, at equilibrium, of leucite crystals and liquid. That 
we have obtained equilibrium is unquestionable, for we 
were able to approach it from opposite directions. 
Whether we started with crystalline potash feldspar or 
with a glass of that composition we always obtained 


8 days 
8 days 


8 days 


5 days 


3 days 


2 hrs. 
20 hrs. 


20 hrs. 
2 hrs. 

Pure sy 


5 days 


3 days 


2 hrs. 
20 hrs. 

8 Morey & Bowen — Melting Potash Feldspar. 

leucite crystals and glass in this 360° temperature-inter- 
val. When held at 1510° for 2 hours glass and small 
crystals of leucite were obtained. When the heating was 
continued at this temperature for 20 hours the leucite 
crystals became larger and less numerous. They evi- 
dently grew freely in equilibrium with the liquid. On 
the other hand when the same material was held at 1535° 
for only 2 hours the crystals completely disappeared. 
There can, therefore, be no question that the persistence 
of crystals at 1510° for 20 hours is the result of the fact 
that they are in equilibrium with the liquid. Incidentally 
it may be noted, too, that, for the attainment of equi- 
librium, prolonged heating is unnecessary at the higher 
temperatures (in the neighborhood of 1500°). 

Incongruent Melting. 

Definite crystalline compounds may melt in one of two 
different ways. A compound of one class melts com- 
pletely at a definite temperature, giving a liquid of the 
same composition as the crystals, for which reason it is 
said to melt congruently. The temperature at which this 
occurs is a true melting point. A compound of the other 
class forms, at a definite temperature, a liquid of different 
composition from its own and at the same time one or 
more new crystalline compounds. Such a compound is 
said to melt incongruently or, sometimes, to melt with 
decomposition. The temperature at which this occurs is 
not a true melting point. Thus the compound CaAl 2 Si 2 O s 
(anorthite) has a true or congruent melting point at 
1550°. At that temperature crystals of the composition 
CaAl 2 Si 2 8 are in equilibrium with a liquid of the same 
composition. It constitutes, by itself, a one-component 

The compound MgSi0 3 is a familiar example of a com- 
pound of the other class. At 1557° the crystalline clino- 
enstatite'(MgSi0 3 ) melts incongruently or breaks up into 
liquid and another crystalline compound Mg 2 Si0 4 (for- 
sterite). There is no temperature at which clinoenstatite 
is in stable equilibrium with a liquid of its own composi- 
tion. It, therefore, has no true melting point. The com- 
pound MgSi0 3 does not constitute, by itself, a one-compo- 
nent system but can be treated only as a part of a 

Morey & Bo wen — Melting Potash Feldspar. 9 

two-component system. In order to represent its behav- 
ior graphically one must construct a two-component 
diagram. Thus fig. 2 represents equilibrium in mixtures 
of Mg 2 Si0 4 and Si0 2 and shows MgSi0 3 as a compound 
of these with an incongruent melting point at 1557°. 
From the diagram we may read off that MgSi0 3 breaks 
up at 1557° into Mg 2 Si0 4 and liquid somewhat more sili- 
ceous than MgSi0 3 . Above 1557° a mass of the compo- 
sition MgSi0 3 is made up of crystals of Mg 2 Si0 4 
(forsterite) and liquid, the forsterite gradually dissolv- 








& \ 











V / 







20 MgS.0.3 40 WT PER CENT 80 

Fk. 2. — Equilibrium diagram of the system Mg 2 Si0 4 — Si0 2 . 

ing as the temperature is raised, until at 1577° the for- 
sterite completely disappears and only then is a liquid 
of the composition MgSi0 3 obtained. 6 

A compound of three oxides may melt incongruently 
in such a way that it breaks up into a liquid and two 
new crystalline compounds. Monticellite (CaMgSiOJ 
shows this behavior. At 1498° it breaks up into Ca 2 Si0 4 , 
MgO, and liquid. It can be treated only as a part of a 
three-component system. 7 However, a compound of 
three oxides, which melts incongruently, does not neces- 
sarily break up in such a way that it must be treated as 
a three-component system. It may form liquid and only 

"For a full discussion see Bowen and Andersen, this Journal, 37, 487-500. 

' Ferguson and Merwin, this Journal, 48, 116. 1919. 

10 Morey & Boiven — Melting Potash Feldspar. 

one new crystal phase, in which case it can be treated as 
a two-component system with relations in every respect 
analogous to those of the binary compound MgSi0 3 . 

"Fig. 3. 





















KAl$i 2 O s io KAl$i 3 8 30 40 50 60 70 80 90 *0 

Fig. 3. — Diagram illustrating the melting of orthoclase. 

The compound, KAlSi 3 8 , with which we are here con- 
cerned, breaks up in this latter manner. At 1170° it 
forms leucite, KAlSi 2 O , and liquid, and its behavior can 
be completely described in terms of a diagram in which 
KAlSi 2 6 and Si0 2 are taken as components, when 
KAlSi 3 8 (orthoclase) becomes a binary compound of 
these. In fig. 3 the general form which this diagram 
must assume is shown. We have accurately determined 
the temperatures on this diagram only at the composition 
KAlSi 3 O s . Work is now being carried out on the whole 

Morey & Bo wen — Melting Potash Feldspar. 11 

system, but it presents many difficulties. On the leucite 
side of orthoclase the temperatures are very high, on 
the silica side the mixtures react very sluggishly and 
require excessively long heating. In the meantime we 
have thought it well to place on record our results on 

The point E (fig. 3) is accurately fixed as is also the 
temperature of the line AB. The composition of the 
liquid B, which is in equilibrium with both orthoclase and 
leucite, we have not yet fixed accurately. The proportion 
of leucite in a mixture of composition KAlSi 3 8 , held at 
about 1250°, is very difficult to estimate, but appears to 
be about 20 per cent. We have placed the point B in 
accordance with this estimate. The melting point of 
leucite we know to be very high, above that of platinum 
(1755 c ). s The melting point of Si0 2 (cristobalite) has 
been fixed at 1710°. 9 Of the temperature and composi- 
tion of the indicated eutectic between orthoclase and Si0 2 
we know nothing. Except at the composition KAlSi 3 O s 
we wish to stress only the general form of the diagram, 
not its details. That it must assume this general form 
cannot be doubted. 

Petrogenic Significance of the Incongruent Melting of 
Orthoclase; by N. L. Bowen. 

The discovery of the fact that orthoclase melts with 
decomposition into leucite and liquid is of great signifi- 
cance to petrogenic theory. It is the first definite exper- 
imental demonstration of a genetic relation between a 
mass consisting of feldspar and feldspathoid and a mass 
consisting of feldspar and free silica, in other words, 
between what may be termed an alkaline and a sub-alka- 
line mass. Moreover, it shows plainly the nature of this 
relation as a fractional crystallization phenomenon. 

Before going into this matter more specifically it is 
desirable to discuss the course of crystallization in mix- 
tures of various compositions shown in fig. 3. 

Reference to that diagram shows that crystallization of 
a liquid of composition KAlSi 3 8 takes place under equi- 

s Leucite crystals were prepared by one of us some years ago when investi- 
gating kaliophilite. To melt the leucite it was found necessary to use an 
iridium container (N. L. Bowen, this Journal, 43, 117. 1917). 

9 Ferguson and Merwin, this Journal, 46, 424. 1918. 

12 Morey & Bowen — Melting Potash Feldspar. 

librium. conditions in the following manner. At 1530° 
■(E) leucite begins to crystallize and it increases in amount 
as the temperature falls, until at 1170° orthoclase begins 
to separate and leucite to redissolve, or, stated better, 
perhaps, the liquid reacts with the leucite, transforming 
it into orthoclase. This continues at constant tempera- 
ture until the liquid and leucite are entirely used up and 
the mass consists entirely of orthoclase. 

A mixture lying on the leucite side of KAlSi 3 8 begins 
to crystallize at a temperature higher than that of the 
point E. In this case when the temperature 1170° is 
reached the reaction between liquid and leucite proceeds 
as before, but all the liquid is used up while still some 
leucite remains and the completely crystalline mass con- 
sists of orthoclase and leucite. 

A mixture on the silica side of orthoclase, if not richer 
in silica than that represented by the point B (probably 
corresponding to about 10 per cent free silica), also 
begins to crystallize with separation of leucite but at a 
temperature lower than that of the point E. At 1170° 
the reaction is completed as a result of the disappearance 
of leucite and the mass now consists of orthoclase and 
liquid of composition B. The liquid then proceeds to 
crystallize along the curve BC, with separation of ortho- 
clase, until at C free silica separates as well and the whole 
mass is now solidified as a mixture of orthoclase and 
free Si0 2 . The exact temperatures and compositions are 
in these parts quite unknown. Only in mixtures with a 
greater excess of free silica than that represented by the 
point B does the early separation of leucite fail. 

Such is the behavior of the various liquids when perfect 
equilibrium obtains. Now in any of these mixtures 
leucite might fail to react completely with the liquid at 
the reaction point (1170°) as a result of the formation 
of an armor of orthoclase about it. The consequence of 
this would be that some liquid of the composition B would 
be left over in all of the mixtures above discussed, even 
that on the leucite side of orthoclase, and this liquid would 
then form orthoclase and free silica. Or if leucite 
crystals were locally segregated that part of the mass 
from which they were removed would crystallize as ortho- 
clase and free silica, even though the composition of the 
original liquid were on the leucite side of orthoclase. 
Conversely, too, even though the composition of the 
original liquid were a moderate distance on the silica side 

Morey & Boiven — Melting Potash Feldspar. IS 

of ortkoclase, the early formed leucite crystals might 
collect locally and if so they would not be completely 
used up by reaction with the liquid and a localized leucite- 
bearing mass would result. 

Xow this early separation of leucite from a mixture of 
the composition of orthoclase cannot be immediately neg- 
atived by the addition of other components. Excess silica 
rapidly neutralizes the tendency, of course, but other sub- 
stances cannot have a comparable effect. Our microcline 
from North Carolina shows that 25 per cent foreign 
material (mainly albite) is insufficient to neutralize this 
tendency, in fact the interval in which leucite appears is 
as much as 270°, so that it would plainly require a con- 
siderably larger amount of albite to bring about the dis- 
appearance of the leucite field. Therefore, in magmas 
rich in orthoclase, and even in those containing mucli 
albite as well, if at the same time they contain not more 
than a small excess of free silica, it is to be expected that 
this early separation of leucite may occur. The leucite 
crystals should disappear with falling temperature but 
the two factors noted in the foregoing may intervene to 
bring about their persistence. The factors are the armor- 
ing of the leucite crystals or their local collection, or 
indeed both. 

Evidence of the Existence of Similar Relations in Natural 
Leucite Rocks. 

Certain described rocks give evidence of the occurrence, 
under natural conditions, of the phenomena noted. 
Hussak describes a rock from Brazil consisting of pheno- 
crysts of leucite (now pseudoleucite) in a groundmass of 
quartz and feldspar, which rock he terms a leucite granite 
porphyry. Evidently strongly influenced by the dictum 
of Rosenbusch and Zirkel that leucite and quartz cannot 
occur together, Hussak suggests the possibility that the 
leucites are remnants of fragments of a foreign rock, 
caught up by the granitic dike. In his final conclusions, 
however, lie inclines towards the opinion that it is an 
ordinary igneous rock from which the mineral combina- 
tion quartz-leucite-orthoclase has crystallized. 10 Our 
results show plainly that such a mineral combination is 
possible, particularly with the relations he notes, namely, 
phenocrysts of leucite in a groundmass of quartz and f eld- 

10 E. Hussak, Xeues Jalirb., 1, 27. 1900. 

14 Morey & Bowen — Melting Potash Feldspar. 

spar. To be sure, if perfect opportunity for reaction 
were presented, either quartz or leucite should be absent, 
but it is easy to conceive of conditions under which such 
opportunity would fail. 

Cross describes a rock from Wyoming which contains 
leucite. Bulk analysis of this rock shows, however, that 
the silica present is entirely adequate to have formed 
orthoclase with the potash and alumina present. It 
seems necessary, therefore, to assume that the glassy 
groundmass is highly siliceous. 11 Such an assumption 
would be entirely justified in the light of our results. 

There is one occurrence that appears to illustrate in a 
convincing manner the separation of leucite under the 
influence of gravity and the consequent formation of a 
quartzose differentiate in those parts from which the 
leucite has moved. Near Loch Borolan in Scotland there 
is a differentiated laccolith which is described by Shand 
as being made up of the following in stratiform arrange- 
ment, as stated below, and shown in fig. 4: 12 

I. Quartz syenites (nordmarkite with 12 per cent, quartz and 

other more quartzose types). 
II. Transition zone of quartz-free syenites. 

III. Feldspathoid-bearing syenites. 

IV. Probable ultra-basic zone (noted in one locality). 

The syenites of III are in part demonstrably pseudo- 
leucite bearing. If it can be imagined that the laccolithic 
chamber was filled with a magma very rich in alkaline 
feldspars, and with not more than a moderate excess of 
free silica, this magma might, as our results show, begin 
to crystallize with separation of leucite. The actual pro- 
portions of the various rock types in the composite mass 
(see fig. 4) show that the general liquid would be very 
rich in feldspar, principally orthoclase, and excessively 
poor in the molecules that go to make up the heavy 
minerals. The density of such a liquid as a glass at 
ordinary temperatures would probably be not far from 
that of rhyolitic obsidian (2.37). 13 The density of leucite 

11 W. Cross, this Journal, 4, 122 and 132. 1897. 

12 S. J; Shand, Trans. Edin. Geol. Soc, 9, pt. Ill, p. 202, 1909, and pt. V, 
p. 376, 1910. See also Home and Teall, Trans. Boy. Soc. Edin., 37 pt. 1, 
p. 163, 1892. 

13 H. S. Washington, Khyolites of Lipari, this Journal, 50, 449. 1920. 

Morey & Boicen- — Melting Potash Feldspar. 
Fig. 4. 



•ij 3 d b ] 

■BuiSjC OU 03 

S 0U 35U 3 


CO « 



- L 

a kqpn 

"cjias^j Jo 8Hi-| 

pcoy - 

ISAiy 3JOUjp3-J 





<u co Tf n o> a o 

~ oi « _a -c CD '^ 
a .O 

F 2: « 

- i 

I 3 O 

B B I 10 

Fig. 4. — Sections of the laccolith at Loch.Borolan, Scotland (after Shand). 

16 Morey & Bowen — Melting Potash Feldspar. 

at ordinary temperatures is 2.46. At higher tempera- 
tures, then, there would probably be a definite though 
small margin of density in favor of leucite crystals, 
especially since the liquid would, no doubt, contain a 
considerable amount of volatile substances, whereas the 
obsidian noted above is nearly free from these. Presum- 
ably the leucite crystals could settle under the influence 
of gravity; indeed, the general arrangement of zones 
can scarcely leave room for doubt that they did and gave 
a lower zone (III) much enriched in that mineral. When 
the time of reaction of the leucite with liquid arrived 
there would be an amount' of leucite above that requisite 
for the reaction and some would be left in excess. The 
excess leucite was, during the further crystallization of 
the mass, transformed into pseudo-leucite (orthoclase + 
nephelite) which is the normal fate of natural leucites 
when cooled slowly under deep-seated conditions. On 
the other hand, in the zone from which the leucites were 
entirely removed (I) there were no crystals to react with 
the liquid (which would be the natural analogue of our 
liquid B, fig. 3) and it crystallized appropriately with an 
excess of free silica. 

We are thus able not merely to accept Shand's inter- 
pretation of the differentiation as gravitative, but we may 
go further and connect the differentiation with the course 
of crystallization ; in short, we may state that the differ- 
entiation was due to fractional crystallization under the 
influence of gravity. One may still accept the possibility 
that the magma was affected by absorption of limestone, 
but the formation of both a quartzose and a f eldspathoid- 
bearing portion from a homogeneous mass shows plainly 
that desilication of feldspar molecules by limestone was 
not essential to the production of feldspathoids. The 
fact that leucite can exist under certain conditions in equi- 
librium with a liquid containing excess silica is the secret 
of the coexistence of two such differentiates. 

Bearing of the Results on the Origin of Nephelite 


Now the rocks of zone III, while undoubtedly at one 
time leucite rocks, in part, at least, are now simply 
nephelite syenites. This is the result of the usual change 
of the leucite to orthoclase and nephelite, already noted. 

Morey & Bowen — Melting Potash Feldspar. 17 

Plainly, then, in the early separation of leucite, itself a 
consequence of the incongruent melting of orthoclase, lies 
the key to the origin of these nephelite syenites. 

This conclusion opens up the whole question as to 
whether many other nephelite syenites may not have been 
formed in a similar way. Leucite may have formed at 
a certain stage and the evidence of its formation, after 
its breakdown into orthoclase and nephelite, may not 
always have been preserved in the form of the pseudo- 
leucite structures. Even in the example described by 
Shancl the nephelite of the pseudo-leucites has suffered 
a further change to muscovite and a zeolite. Through 
such further changes, not necessarily exactly of this kind, 
the evidence of the former existence of leucite may fre- 
quently, perhaps, have been entirely destroyed. More- 
over, upon the change of leucite into orthoclase and 
nephelite the course of crystallization will follow lines 
upon which our work throws no light and it may be 
possible that even the more common, highly sodic, 
nephelite syenites could form as further differentiates. 

Examining nephelite syenites and their associates .with 
these considerations in mind, we find a considerable 
amount of evidence that the formation of leucite may have 
occurred as an intermediate step in their genesis. 

The Himausak batholith in Greenland is stratified in 
a manner closely related to that shown by the Loch 
Borolan mass. At the top is a quartzose phase, arfved- 
sonite granite, which passes downward into quartz-free 
syenite and finally into sodalite and nephelite syenites of 
great variety. 14 In one of these, the so-called lujavrites, 
which are the lowest exposed rocks of the mass, there 
are large crystals of analcite which Us sing presents 
reasons for believing were formerly leucite. 15 The 
nature of the stratification and the existence of these 
pseudo-leucites (?) is so strikingly similar to the condi- 
tions at Loch Borolan that one must consider the possi- 
bility that the early separation of leucite has been a 
factor controlling the differentiation of the mass (see 
fig. 5). 

14 X. V. Ussing, Geology of the Country around Julianehaab, Greenland, 
Med. om Gronland, vol. 28, p. 322, fig. 29,' 1911. 

15 Op. eit. pp. 164, 165. 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 19.— July, 1922. 

18 Morey & Bowers— Melting Potash Feldspar. 

The igneous massif of Bezavona in Madagascar , as 
described by Lacroix, consists of quartz syenites, nephe- 
lite syenites, monzonites, and other types, including vari- 
ous dike and flow rocks. In the coarse granular rocks 
there is apparently nothing suggesting the formation of 
leucite, but in quickly chilled facies, which may be 
regarded as quenched, leucite appears. Thus there are 
microsyenites with leucite and also leucite phonolites. 16 
These facts again suggest the possibility that the forma- 
tion of leucite may have been an intermediate step in the 
genesis of the nephelite rocks. 

Fig. 5. 

Lujavrites "-—'-.-—-.-. .,.--.------ g Kakor tok i fcfi , 

Fig. 5. — Section of the Ilimausak mass, Greenland (afer Ussing). 

Still less definite as evidence, and yet worthy of consid- 
eration, are certain structures observed in nephelite rocks 
that suggest the former presence of leucite. In some 
members of the Ice River complex of British Columbia 
there are spots of a "finger-print-like" intergrowth 17 of 
orthoclase and nephelite that is practically identical with 
the "dactylotype" intergrowth of these minerals when 
they form pseudo-leucites. Structures that are perhaps 
of similar origin are described by Lacroix from the nephe- 
lite syenites of the Los Archipelago 18 and by Brouwer in 
rocks from the Transvaal. 19 

16 A. Lacroix, Les Boches Alcalines d 'Ampasindava, Nouv. arch, du 
museum, Paris. Serie 4, 5, 197 and 207, 1903. 

17 J. A. Allan, Geology of the Field Map-area, B. C. and Alberta. Geol. 
Surv. Han., Mem. 55, p. 133 and p. 285, 1914, Plate XVIIB. 

13 A. Lacroix, Les syenites nepheliniques de 1 'archipel de Los. Nouv. arch, 
du museum, Serie 5, 3, p. 53, 1911. 

19 H. A. Brouwer, Transvaal Nephelien-Syenieten. p. 40, PL I, Fig. 1. 

Morey & Boiven — Melting Potash Feldspar. 19 

An intimate relation between lencite and nephelite 
rocks is observed in many fields. A striking example is 
shown by the Magnet Cove complex. In particular it 
is noteworthy that a lencite porphyry and a f oyaite, asso- 
ciated there, have nearly identical chemical composition, 
at least in some specimens. Washington has called atten- 
tion, also, to a similar relationship of a lencite-rich rock 
(lencite phonolite) of the Sabatinian district, Italy, with 
a nephelite syenite of Beemerville, N. J. 20 

In volcanic fields the frequent intimate association of 
trachyte, leucite-trachyte and phonolite is suggestive in 
this connection. We may mention only the Laacher See 
area from which region came one of the feldspars (sani- 
dine) which we used in our investigation and which shows 
in a typical way the decomposition into lencite and liquid. 

On the whole, then, there is a considerable body of evi- 
dence pointing to the importance of the incongruerit melt- 
ing of orthoclase as a factor not merely in the formation 
of leucite rocks but of other feldspathoid-bearing rocks 
as well. It is not intended here to set this up as the sole 
factor involved in the formation of feldspathoid rocks. 
There are many indications that some leucitic rocks are 
formed as a result of differentiation along lines which 
produce a liquid rich in mica molecules and that then the 
extrusion of the liquid places it in surroundings where 
it is unable to retain the water necessary for the forma- 
tion of mica, with the result that leucite is formed just 
as it is when mica is melted in an open crucible. 21 The 
more basic leucite rocks, leucite basalts, etc. may perhaps 
have formed in some such way. 

In former papers the writer has presented reasons for 
believing that in the reactions which are revealed by the 
presence, side by side in a rock, of alkaline feldspar, mica, 
and quartz, there is evidence of the breakdown of the 
polysilicate (feldspar) molecules into less siliceous 
(feldspathoid) molecules and free silica. This was 
believed to be due to the hydrolyzing action of water in 
the magma. 22 It was not then anticipated that it would 
be found that potash feldspar breaks down in a similar 

20 H. S. Washington, Igneous complex of Magnet Cove, Arkansas, Bull. 
Geol. Soc. Amer. 11, 399, 1900; and Roman Comagmatie Eegion, Carnegie 
Institution of Washington, Pub. No. 57, 1906, p. 47. 

21 H. S. Washington, The formation of leucite in igneous rocks. J. Geol., 
15, 379, 1907; and N. L. Bowen, The later stages of the evolution of the 
igneous rocks, J. Geol. Suppl. to vol. 23, 60, 1915. 

20 Morey & Bowen — Melting Potash Feldspar. 

way even in the dry melt. The discovery of this fact 
should not, however, lead lis to reject the effects of water 
as important in promoting such behavior of the polysili- 
cates. 23 Nor need we abandon the suggestion that in the 
preparation in the liquid of f eldspathoid molecules, under 
the influence of water, there enters the possibility, with 
appropriate fractional crystallization, of the formation 
of feldspathoid-bearing rocks. However, these reactions 
represent, as the writer has always admitted, 24 a consid- 
erable extrapolation from any facts that have yet received 
laboratory demonstration. It is with considerable satis- 
faction, therefore, that we announce a laboratory demon- 
stration of the fact that a mass consisting in one part of 
feldspar and quartz and in another of feldspar and feld- 
spathoid can form from a single homogeneous liquid. 
The method of formation of these contrasted parts, which 
may be referred to as subalkaline and alkaline, respec- 
tively, is the method of fractional crystallization, which 
has also been shown fairly definitely to be adequate for 
the production of all varieties of subalkaline rocks from 
one liquid. 25 


A pure synthetic orthoclase was prepared by crystal- 
lizing glass of the composition KAlSi 3 8 in a bomb with 
water vapor. This material is particularly suitable for 
the determination of the melting-point of pure orthoclase 
and was used for that purpose. The temperature ordi- 
narily given as the melting-point of orthoclase is about 
1200° and has been determined on natural crystals. 
When our artificial crystals were held at 1200° for a week 
they gave a product which had the appearance of a glass, 
megascopically, but which, examined under the micro- 
scope, showed a structure described as a very fine cross- 
lining. At higher temperatures this structure became 
more distinct, taking the successive forms shown in iig. 1, 

22 N. L. Bowen, The later stages of the evolution of the igneous rocks, J. 
Geol., Suppl. to vol. 23, 60. 1915. 

23 In dry melts albite shows no tendency to decompose in a manner analo- 
gous to that shown by orthoclase. 

24 N. L. Bowen, Crystallization-differentiation in magmas, J. Geol. 27, 395. 

25 N. L. Bowen, The later stages of the evolution of the igneous rocks, 
J. Geol., Suppl. to vol. 23. 1915. 

Morey & Boiuen — Melting Potash Feldspar. 21 

and finally becoming' typical lencite crystals. The point 
at abont ±200° is therefore not the true melting-point of 
orthoclase bnt is the temperature at which it melts incon- 
gruently, breaking np into liquid and lencite. The exact 
temperature of this decomposition we have determined 
as somewhat lower than 1200°, namely, about 1170°. The 
temperature of final disappearance of leucite is about 
1530% so that the interval of incongruent melting is 
remarkably large, viz., 360°. Three natural potash feld- 
spars, microcline from North Carolina, sanidine from 
Laacher See, and adularia from St. Gotthard show the 
same kind of behavior, though in these the upper limit 
of melting (disappearance of leucite) is lowered some- 
what through the presence of foreign matter. 

This incongruent melting of orthoclase is of particular 
importance in petrogenic theory because it shows plainly 
how, by fractional crystallization, a homogeneous liquid 
could form a differentiated mass consisting of orthoclase 
and leucite in one part and of orthoclase and free silica 
in another. It shows, too, that leucite can form from a 
liquid containing an adequate amount of silica to form 
orthoclase and that a mass may have leucite as early 
crystals (phenocrysts) together with free silica as late 
crystals (groundmass). These considerations explain 
the occurrence of such a rock as the leucite granite por- 
phyry of Brazil and such a differentiated mass as the 
syenite laccolith at Loch Borolan, Scotland. It is to be 
noted that both these occurrences show pseudo-leucites, 
formed secondarily after leucite, and consisting, as do 
the leucite s of intrusive rocks in general, of an inter- 
growth of orthoclase and nephelite (or secondary 
products after nephelite). This regular behavior of leu- 
cite in breaking up into orthoclase and nephelite suggests 
that the early separation of leucite, with a subsequent 
change of that nature, may afford a key to the origin of 
many nephelite rocks as well as leucite rocks. 

Geophysical Laboratory, 

Carnegie Institution of Washington, 
Washington, D. C, 
March, 1922. 

22 F. von Huene — Order Thecodontia. 

Akt. II. — The Triassic Reptilian Order Thecodontia; by 
F. von Huene. 

During- the last several years the writer has been much 
occupied with reptiles of the order Thecodontia (see Nos. 
10-20 of the literature list at the end of this paper) and 
allied groups. In the present paper I am going to give 
briefly the results as regards classification and relation- 
ship. The latest literature is given at the end, and all 
other papers will be found quoted in these. 

The order Thecodontia (E. Owen 1859) consists of 
three suborders: Pseudosuchia (Zittel 1889), Parasuchia 
(Huxley 1875) and Pelycosimia (Huene 1911). The 
animals constituting these three suborders are of very 
dissimilar form and size, but are anatomically very nearly 
related. The Pseudosuchia form the radicle stock of the 
whole group. Both of the other suborders spring from 
early Pseudosuchians, but have no descendants them- 
selves; the Pseudosuchians give rise probably to all 

The Pseudosuchia I propose to classify as follows : 

(Proterosuchus fergusi 

Proterosuchidse . . . \Dyoplax arenaceus 

[Erpetosuchus granti 

Sphenosuchida? Sphenosachus acutus 

Ornithosuchus woodwardi 
Omithosuchus taylori 
Saltoposuchus connectens 
Saltoposuchus longipes 
Pedeticossaurus leviseuri 
ScleromochlidEe Scleromochlus taylori 

Euparkeriidse . . S Eu P ar ^ er i a capensis 

' VBrowniella afficana 

Aetosauridffi Uetosaurus ferratus 

(Aetosaurus crassicauda 
Stegomosuchidse Stegomosuchus longipes 


With regard to the last of these forms, it was first 
described as Stegomus longipes by Emerson and Loomis. 6 
Then the writer re-investigated it at Amherst in 1911 and 
published his results in 1914. 13 The skull now agreed 
with some of the other Pseudosuchians, but extremities 
and dermal plates were different. It has a long skull and 
less than half of its length is preserved. This form can- 

F. von Huene — Order Thecodontia. 23 

not go generically with the older animal described by 
Marsh as Stegomus arcuatus (see also 13 ), which I now 
take for a primitive Parasuchian. Therefore I propose 
to call the former Stegomosuchus (n. gen.) longipes and 
its Psendosuchian family Stegomosuchidse (n. fam.). 

The classification of the Parasuchia is mostly based 
npon features of the skull. The essential points are: 
relative length of the base of the skull, relative length of 
the snout, position of the narial openings, condition of 
the supratemporal opening, and palate. The posterior 
part of the skull (beginning in front with the anterior 
margin of the nares) in different genera has a relative 
length of from 48 to 33.3 per cent of the whole skull. 
In some very primitive genera, however, it cannot yet 
be measured as the tip of the snout is missing in the 
known specimens. 

The Parasuchia may be classified in the following 
manner : 

Desraatosucliidag Desmatosuclius 


Stagonolepidae -j Stagonolepis 

{ ? Stegomus ( arcuatus ) 
Phytosauridse Phytosaurus 





Wpiscoposaurus ' 


Riley a 

Angistorhinopsis, n. gen. 


The Desmatosuchidae and the Stagonolepidas I regard 
as the most primitive families, of not later than Middle 
Triassic age. The European f Phytosaurus I take as 
a persistently primitive form retaining an early stage of 
Parasuchian evolution in the very long posterior part of 
the skull and the dermal armature. But the shifting 
backward of the supratemporal groove and the short base 
of the skull nevertheless indicate a terminal member of 
this branch of the Parasuchia. The Mystriosuchidse are 
a big group which probably in the future will be divided 
into at least two families, as their feet show very different 



F. von Huene — Order Thecodontia. 

structures, but the evidence is not yet complete enough 
to do this. Further, it might be noted that "Rutiodon" 
manhattanensis probably does not belong to this genus 
but to another. 

In a paper still in press 20 an extensive discussion is 
given of the history of the Parasuchia, and in another 
a general view of the Thecodontia. 

The writer holds 13 that the Pseudosuchia give rise to 
the Archosauria. The reasons for this need not be 
repeated here. From forms probably nearly related to 
the Ornithosuchidse, the Ornithischia and the Aves prob- 
ably arose through adaptations and the Pterosauria not 
very far from them. The Crocodilia also probably came 
from that part of the stem. But the Saurischia the 
writer takes to be an offshoot of the very earliest Pseudo- 
suchians in the most ancient Trias sic time. 


/^p^oJiTTac^e Acrosaundcte jSphenoc/orttictae 
<$> ■& £copodjLia,__ 

der''*'**^ 4.Kfl 


■ In 1920 17 the writer expressed the opinion that the 
Khynchocephalia (with the taxonomic rank of an order) 
are descendants of the same root as the Thecodontia. 
If that is true, it would be easy to understand why so 
many characters are common to both phyla. From this 
viewpoint, the Gnathodontidas (Howesia, Mesosuclius, 
Br achy rhino don, Polysphenodon and probably Eifelosau- 
rus) would form the most primitive family of the Rhyn- 
chocephalia. The contemporaneous family Rhyncho- 
sauridse (Ilhynchosaurus , Hyperodapedon and Sten- 
ometopon) is little more specialized. The stem of the 
Rhynchocephalia is represented in later times by the 
AcrosauridsB in the Upper Jurassic and by the Tertiary 
and present Sphenodontidas. In the Upper Jurassic the 

F. von Huene — Order Thecodontia. 25 

Sauranodontida?, and in the uppermost Cretaceous the 
Champ sosaurida?, branched off from the main line. 

As an Upper Permian Thecodont Broom has described 4 
the genus Youngina. But this form seems to the writer 
very nearly related to Broomia Watson. 27 Watson has 
pointed out that Broomia is nearly related to Heleosaurus 
and Heleophilus. They are also allied with Adelosaurus, 
Aphelosaurus and even with the much more specialized 
Protorosaurus, further with "Eosuchus" (Watson = 
Noteosuclius Broom). All of these genera should appar- 
ently be united in a single inclusive group, the Protoro- 
sauria. Watson has pointed out 27 that Broomia possibly 
might be related to the Lower Permian Captorhinidae, 
and through these to the more typical Cotylosaurians. 
If this chain of connections be true, the Protorosauria 
would form an intermediate link between a group of the 
primitive Cotylosaurians and the Thecodonts, or, in gen- 
eral, the Archosauria. 

Tubingen, 7. January 1922. 


1 Broom, E. : On a new reptile (Proterosuclius fergusi) from the Karroo 
beds of Tarkastad. Ann. S. Afr. Mus., 4, 1904, 159-163, pi. 19. 

2 Broom, E. : The South African diaptosaurian reptile Howesia. Proe. 
Zool. Soc. London, 1906, 591-600, pis. 40-41. 

3 Broom, E. : On the South African Pseudosuchian Euparkeria and allied 
genera. Ibid., 1913, 619-633, pis. 75-79. 

4 Broom, E. : A new Thecodont reptile. Ibid., 1914, 1072-1077, 3 figs. 

5 Case, E. C. : Preliminary description of a new suborder of Phytosaurian 
reptiles, with a description of a new species of Phytosaurus. Journ. of Geol., 
28, 1920, 524-535. 

6 Emerson, B. K. and Loomis, F. B. ; Stegomus longipes, a new reptile 
from the Triassic sandstones of the Connecticut Valley. This Journal (4), 
17, 1904, 377-380, pi. 22. 

T Haughton, S. H. : A new Thecodont from the Stormberg beds. Ann. 
S. Afr. Mus., 12, 1915, 98-105, 3 figs. 

s Haughton, S. H. : On the reptilian genera Euparkeria Broom and Meso- 
suchus Watson. Trans. E. Soc, S. Afr., 10, 2, 1921, 81-88, pis. 2-3. 

9 Hoepen, E. C. 1ST. van: A new Pseudosuchian from the Stormberg beds. 
Ann. S. Afr. Mus., 5, 1915, 83-87, pis. 13-14. 

10 Huene, F. v. : Ueber einen echten Bhynchocephalen aus der Trias von 
Elgin, Brachyrhinodon taylori. N. Jahrb.'f. Min., etc., 1910, 2, 29-62. 

11 Huene, F. v.: Ueber Erythrosuchus, Vertreter der neuen Beptil-Ordnung 
Pelycosimia. Geol. u. Pal. Abh., 10, 1, 1911. 1-60. Tf. 1-11. 

12 Huene, F. v. : Beitrage zur Kenntnis u. Beurteilung der Parasuchier. 
Ibid., 10, 1, 1911, 61-122, Tf. 12-17. 

13 Huene, F. v. : Beitrage zur Geschichte der Archosaurier. Ibid., 13, 1, 
1914, 1-53, Tf. 1-7. 

26 F. von Huene — Order Thecodontia. 

14 Huene, F. v. : On reptiles of the New Mexican Trias in the Cope Collec- 
tion. Bull. Amer. Mus. Nat. Hist., 34, 1915, 485-507, 64 figs. 

15 Huene, F. v. : Bemerkungen zur Systematik u. Stammesgeschichte 
einiger Eeptilien. Zeitschr. f. indukt. Abstammungs- u. A^ererbungslehre, 
22, 3, 1920, 209-212. 

16 Huene, F. v. : Ergebnisse einiger stammesgeschichtlicher Untersuchun- 
gen an fossilen Eeptilien. Ibid., 24, 1920, 160-163, Tf. 7. 

17 Huene, F. v.: Die Osteologie von Aetosauruq ferratus. Acta Zoologica, 
1, 1920, 465-491. 

1S Huene, F. v. : Ein Parasuchier im oberen Muschelkalk von Bayreuth. 
Senckenbergiana, 2, 5, 1920, 143-145, 2 figs. 

19 Huene, F. v. : Neue Pseudosuchier u. Ccelurosaurier aus dem Wiirttem- 
bergischen Keuper. Acta Zoologica, 2, 1921, 329-403, 4 Tf. 

20 Huene, F. v. : Neue Beitrage zur Kenntnis der Parasuchier. Abh. 
Preuss. Geol. Landesanst. (Still in press.) 

21 Jaekel, O. : Ueber einen neuen Belodonten aus dem Buntsandstein von 
Bernburg. Sitz. ber. Ges. nat. Freunde, Berlin, 1910, 197-229. 

22 Lees, J. H. : The skull of Palceorliinus. Journ. of Geol., 15, 1907, 121- 

23 Lull, E. S. : Triassic life of the Connecticut Valley. Conn. Geol. and 
Nat. Hist. Surv., Bull. 24, 1915. 

24 MacGregor, J. H. : The Phytosauria, with especial reference to Mystrio- 
suchus and Ehytidodon. Mem. Amer. Mus. Nat. Hist., 9, 1906, 29-101, 
pis. 6-11. 

25 Mehl, M. G. : The Phytosauria of the Trias. Journ. of Geol., 23, 1915, 
129-165. 20 figs. 

26 Mehl, M. G. : The Triassic fossil bearing horizons near "Wingate, New 
Mexico, with a description of Acompsosaurus wingatensis. Bull. Univ. Okla- 
homa, Studies, 5, 1916, 29-39, 1 pi. 

27 Watson, I). M. S. : Broomia perplexa gen. et sp. nov., a fossil reptile 
from South Africa. Proc. Zool. Soc. London, 1914, 995-1010, pi. 6. 

H. L. Wells — Discussion of Triple Salts. 27 

Art. III. — A Discussion of Triple Salts; by Horace 

L. \Yells. 

[Contribution from the Sheffield Chemical Laboratory of Yale University.] 

The object of this article is to present a few points in 
regard to triple salts, particularly in connection with the 
regularity and irregularity of their types, without 
attempting to give a complete list of those that are known, 
or to discuss them fully. 

There are many instances where analogous triple salts 
are known, and in some cases these occur in rather exten- 
sive series, but there are a great many cases where anal- 
ogy is lacking between salts of analogous metals, so that 
there appear to be no definite laws, based upon the 
valency or other characters of the constituent salts, 
according to which they appear to be formed. A similar 
conclusion was reached by the writer 1 in connection with 
a discussion of double halogen salts. 

To give examples, the triple chlorides recently 
described by the writer, 2 

Cs 4 Ag 2 Au 2 Cl 12 , Cs 4 Au' 2 Au"' 2 Cl 12j 

Cs 4 ZilAti 2 C1 123 Cs 4 HgAu 2 Cl 12 , 

Cs 4 CtiAu 2 C1 12 , 

show analogous formulas and isomorphism, even where 
two atoms of a univalent element and one of a bivalent 
element replace each other, but Pollard's salt, 3 (XHJ 6 Ag 2 
Au 3 Cl 17 , fails to agree with them, as do also two triple 

KFe"Fe'",Br .3H,,0 and RbFe' / Fe" , 2 Br 9 .3H 2 0, 

described in this laboratory by Professor P. T. vTalden, 4 
although, in this case, there are agreements in the valen- 
cies of the metals. 

Among the considerable number of triple thiocyanates 
described by the writer and his associates, 5 many analo- 
gous salts were found, but the number of types that 

1 Amer. Cliem. Jour., 26, 389. 

2 This Journal, Mav, 1922. 

3 This Journal, April. 1922. 

4 This Journal, 48, 283, 1894. 

5 H. L. Wells, O. G. Hupfel, H. F. Merriam, C. S. Leavenworth, R. T. 
Roberts, Amer. Chem. Jour., 28, 245; P. L. Shimi and H. L. Wells, Ibid., 
29, 474; H. L. Wells, Ibid., 30, 144.] 

28 H. L. Wells — Discussion of Triple Salts. 

occurred with metals of corresponding valency is remark- 
able. A list of these thiocyanates, according to their 
types, is as follows : 

2:1:2 salts. 
Cs 9 MgAg 2 (SCN) 6 .2H 9 Cs 2 CaAg 2 (SON) 6 .2H,0 

Cs 9 MnAg 9 (SCN) 6 .2H.,0 Cs„NiAg 2 (SCN) 6 .2H,0 

Cs 2 CoAg 9 ( SCN) 6 .2H o 0' Cs 2 NiCu' 2 (SCN) 6 .2H 2 

Cs 9 CdAg 9 (SCN) 6 .2H o RboBaAg",(SCN) 6 .2H 2 

Cs 2 CdAg 2 (SCN) 6 

3:1:2 salts. 
Cs 3 SrAg 9 (SCN) 7 Cs 8 BaAg„(SCN) 7 

Cs 3 SrCu' 2 (SCN) 7 Cs 3 BaCu' 2 (SCN) T 

4:1:2 salts. 
K 4 BaAg 2 (SCN) 8 .H 2 Rb 4 BaAg 2 (SCN) s .H 2 

1:1:1 salt. 2:1:1 salt. 

CsZnAg ( SCN ) 4 .H 2 Cs 2 ZnAg ( SCN) 5 

1 • 1 • 3 salt I • 2 • 3 salt 

CsCdAg 3 (SCN) 6 .2H 2 6 CsZn 2 Ag 3 (SCN) 8 

2:1:4 salt. 1:2:4 salt. 

Cs 2 CdAg 4 ( SCN) 8 .2H 2 CsZn 2 Ag 4 ( SCN) 9 

In most cases only a single triple salt could be pre- 
pared from the same three simple thiocyanates, but Rb, 
Ba,Ag gave two of them, one of which is analogous to 
the K,Ba,Ag salt, while neither of them corresponds to 
the single Cs,Ba,Ag compound. With Cs,Zn,Ag and Cs, 
Cd,Ag four different salts were prepared in each case, 
but there is no correspondence between these compounds 
of such closely related metals as zinc and cadmium. It 
is a curious circumstance also that while two of the cad- 
mium salts, as hydrous and anhydrous forms, belong to 
the most common, 2:1:2, type, the other two cadmium 
compounds and all four of the zinc salts are unique in 
type, and thus comprise six out of the nine kinds of 
formulas shown in the above table. 

A series of analogous triple nitrites was described in 
this laboratory by Professor George S. Jamieson. 7 
Their formulas are as follows : 

6 Described as Ca t Cd 8 Ag 1 o(SCN)2o.6H 2 0, which varies but little from the 
simpler formula given here. 

7 Amer. Chem. Jour., 38, 614. 

H. L. Wells— Discussion of Triple Salts. 29 

Cs 3 BaAg (NO, ) 6 .2H 2 Cs 3 PbAg (NO, ) 6 .2H,0 

Cs 3 SrAg (N0 2 ) G .2H 2 K 8 PbAg (N0 2 ) 6 .2H 2 

It is to be observed that the compounds in the first 
column contain the same metals as two of the triple thio- 
cyanates, but the formulas do not correspond, nor was 
the 3:1:1 ratio, which these nitrites show, found at all 
among the triple thiocyanates. Triple nitrites appear to 
be formed with particular facility, and a few others will 
be mentioned here to show their types. 

The salt K 2 CaNi(N0 2 ) 6 was described by Erdmann, 8 
and the analogous strontium and barium compounds are 
known also. The compound K 2 NaCo"'(N0 2 ) 6 is well 
known as the precipitate obtained in a qualitative test 
for potassium, while the analogous salts (NH 4 ) 2 NaBi 
(N0 2 ) 6 , Bb 2 NaBi(N0 2 ) 6 and Cs 2 NaBi(N0 2 ) 6 have been 
described by Ball. 9 

A triple chloride described by BonsdorfT 10 a very long 
time ago is worth mentioning on account of its irregular, 
complex formula, K 6 CuHg 3 Cl 14 .2H 2 0. Its composition 
was confirmed in this laboratory by H. F. Merriam, 11 who 
prepared and partially analyzed it. 

Triple cyanides are known, such as the precipitate, 
K 2 CaFe(CN) 6 .3H 2 0, obtained with potassium ferrocy- 
anide in Baubigny's test for calcium. Of course this 
may be called a double ferrocyanide. 

The following are examples of triple salts that occur 
as minerals : 

Paclmolite, NaCaAlF 6 .H,0 
Polvkalite, K,Ca,Mg(S0 4 ),.2H,0 
Thaumasite, CaC0 3 .CaSi0 3 .CaS0 4 .15H,0 
Sulphohalite, 2Na,S0 4 .NaCl.NaF. 
Hanksite, 9Na 2 S0 4 .2Na 2 C0 3 .KCl 
Kainite, K,SO,.MgS0 4 .MgCl,.6H,0 
Northuptite, MgC0 3 .Na,C0 3 .NaCl. 

The first two of these show that fluorides and sulphates 
are capable of forming triple salts, the next two are triple 

s J. prdkt. Chem., 97, 395. 

9 J. Chem. Soc, 87, 761; 95, 2126. To the cesium salt, which is obtained 
as a precipitate in Ball's test for sodium, the formula Cs Na Bi 3 (NO 2 ) 30 was 
ascribed, but the simpler formula, differing but slightly from it and analo- 
gous to the others, appears preferable after a careful consideration of the 
original description. 

10 Pogg. Ann., 33, 81, 1834. 

n Amer. Chem. Jour., 28, 256 (1902). 

30 H. L. Wells — Discussion of Triple Salts. 

salts of single metals with three acids, while the last 
three show combinations with two metals and three acids, 
and with two of each. 

Enough examples have been given to show that triple 
salts of analogous metals may have corresponding formu- 
las in some cases, but that there are many irregularities, 
leading to a large variety of formulas. Some of the 
variations depend, as is the case with the Cs-Zn-Ag and 
Cs-Cd-Ag thiocyanates, upon the proportions of the three 
simple salts in the solutions from which they are depos- 
ited, but this appears to be exceptional, since in most 
cases only a single triple salt can be obtained under such 

A great many of the triple salts show simple numerical 
ratios in their formulas, but a few of them have undoubt- 
edly complex compositions. The frequent occurrence in 
the formulas of 6 and 12 negative atoms or radicals seems 
worthy of mention as suggesting related molecular struc- 
tures in such compounds, but there are many varieties 
of formulas not corresponding to these numbers. 

New Haven, Conn., 

E. L. Troxell — Homed Eocene Ungulates. 


Art. IV. — Horned Eocene Ungidates; by Edward L. 


[Contributions from the Othniel Charles Marsh Publication Fund, Peabody 
Museum, Yale University, New Haven, Conn.] 

Iii the Eocene period two distinct forms of rhinoceros- 
like animals are fonnd to have rugose and thickened 
nasal bones which appear to be the beginning of horn 
supports analogous to those of our modern rhinoceros. 
The first of these is the well known Colonoceras agrestis 
Marsh; the second is Metahyrachyus bicornutus, new 
genus and species. 

In Oligocene time we see nothing again of horned 
rhinoceroses until late in the period, when the dicera- 
theres come in, and one concludes therefore that we 
either have no record of the intervening members of one 
single, great, and continuous race, or that the earlier 
branch was cut off and that nature wrought the later 
forms from another, a hornless group leading through 
Trigonias Lucas and Ccenopus Cope. 

Colonoceras agrestis Marsh. 

(Figs. 1-3.) 
Holotype, Cat. No. 11082, Y. P. M. Eocene (Bridger), near Fort Bridger, 

11082 TYPE 

Y. P. M. 

Fig. 1. — Colonoceras agrestis Marsh. A small specimen closely related to 
Hyrachyus, but having rugosities on the nasals. Eocene (Bridger). X 1 /^- 

The original description of this species is as follows : ] 

1 O. C. Marsh, This Journal (3), 5, 407, 1873. 


E. L. Troxell — Horned Eocene Ungulates. 

"In its cranial characters and dentition, this genus resembles 
most nearly Hyrachyus Leidy, and Helaletes Marsh. It differs 
especially from these genera, so far as they are known, in the 
presence of a pair of dermal horns on the nasal bones, which were 
strengthened to support them. These horns were placed opposite 
each other, and their position, in a nearly perfect skull in the 
Yale Mnseum, is indicated by two rugosities, which have their 
surfaces marked by radiating lines. In the present species, 
which was about as large as a sheep, the horns were widely 

110 82 TYPE 
Y. P M 



Space occupied by seven teeth in upper molar series 77 

Extent of three true molars 41 

Distance between orbits 62 

Distance between apices of horn rugosities 27 

Length of frontals on median suture . 62 

Expanse of occipital condyles 40 

"The remains of this species at present known are from the 
Eocene of Wyoming. ' ' 

So far as one may judge from the teeth, the species 
here described by Marsh is very close to Hyrachyus 
affinis. There are slight variations in the form of the 
third npper molar, but the premolars are of the typical 
Hyrachyus sort, with the strong anterior cross crests, 
the protoloph, enveloping the thin short metaloph which 
lies transversely across the center of the tooth. 

These contrasts may be made : 

E. L. Troxell — Horned Eocene Ungulates. 33 

Colonoceras agrestis 
Strong ribs on metacone of 

M 3 rounded, ant. and post. 

sides not parallel. 
M 3 heavy projecting post, cin- 

Transverse valley of molars 

not blocked. 
Peculiar metastvie on molars. 

Hyrachyus affinis 

Ribs faint or absent. 

M 3 subquadrate, ant. and post. 

sides straight. 
Short cingulum, not extended 

Cingula across valleys. 

Metastvie inconspicuous. 

11082 TYPE 
Y. P. M 

Fig. 3. — Colonoceras agrestis Marsh. X%- 

As compared with the more complete skull of H. affinis 
gracilis nobis, Cat. No. 11170, Y. P. M., one sees these 
differences : C. agrestis has more slender condyles, 
broader frontals, especially between the orbits, and 
finally has the incipient horn rugosities as the generic 
name implies. 

Except for the presence of horns, none of these features 
separates C. agrestis widely from H. affinis, and only 
in a most rigid splitting of groups can the two be distin- 
guished generically. It has been suggested that the 
horns of the early rhinoceroses indicate sex distinctions, 
and this may be true in the present case. 

M et airy r achy 'lis bicornutus, gen. et sp. nov. 

(Figs. 4, 5.) 
Holotype, Cat. Xo. 10258, Y. P. M. Eocene (Bridger), Millersville, near 
Fort Bridger, Wyoming. 

One of the rarer skulls of early rhinoceroides in the 
collection is this specimen found by R. E. Son in 1873. 

Am. Jour. Sci. — Fifth Series, Vol. IV. No. 19.— July, 1922. 


E. L. Troxell — Horned Eocene Ungulates. 

The first incisor is absent, otherwise every tooth of the 
superior series is represented. The lower jaws are not 

In two important respects this specimen is different 
from any Hyrachyus. It has the double internal cone 
on the third and fourth premolars, and has incipient 
horn rugosities on the nasals. 

10258 TYPE 

Y. P. M. 

Fig. 4. — Metahyracliyus bicornutus, gen. et sp. nov. Top view of skull 
showing what appears to be the beginning of the horn supports in the great 
family of the rhinoceroses. X%- 

The first feature, the separation of the tetartocone 
from the deuterocone, at once reminds us of many 
species of the later rhinoceroses, in which there is a 
tendency for the premolars to become molariform. 
Because the metaloph is so abbreviated, we find its 
nearest comparison with Ccenopus (see C. platycephalus 
(Osborn and Wortman) ). 

The second important feature, the appearance of horn 
rugosities on the ^asals of this species, is a matter of 
great significance; for some such race must have given 
rise to Diceratherium armatum Marsh and to Menoceras 
coohi (Peterson) in the later Oligocene and Miocene 
periods, in which we see the same transverse arrange- 
ment of the two elements. This may mark the beginning 

E. L. Troxell — Homed Eocene Ungulates. 35 

of the horn-supports seen in the modern rhinoceroses, 
bnt here the horns, when there are two of them, are placed 
one in front of the other. 

It is well known that Colonoceras agrestis had similar 
thickenings of the nasal bones, even more rugose ; they 
probably had a common ancestry, but that species is far 
separated from the present one in other respects : C. 
agrestis is about two thirds the size; the premolars have 
the simple internal cone ; the strong transverse ridge on 
F 3 is anterior and not medial in position ; and finally the 
rugose areas are midway on the nasals and not posterior 
as in the new species. 

Detailed Morphology of M. bicomutus. — The upper 
incisors increase in size from' front to rear, and there is 
a progressively larger space between each and between 
the last incisor and the canine. The third incisor is 
subcaniniform, being rather long and slightly recurved, 
but it is narrow transversely. The canine is only 
moderately long; it is compressed and recurved and 
therefore bears no resemblance to those of the later 
rhinoceroses except Hyracodon Leidy. 

The first premolar, although elongated fore and aft, 
is much broader than that in Hyrachyus. This tooth 
has the one main protocone and on the inner side an inci- 
pient deuterocone with minor ridges. The second pre- 
molar is nearly circular in form, consisting mostly of 
the large strong outer cone (protocone), with a small 
tritocone behind it, and of one prominent inner cone, the 
deuterocone. A thin ridge runs from the latter to the 
middle of the ectoloph and divides the tooth equally; 
this is a distinctive feature of the genus. 

Undoubtedly the most important feature of the skull 
is the double inner cone on each of the larger premolars. 
It shows the beginning of the separation of the tetarto- 
cone from the deuterocone, and a first step toward the 
assumption of the molariform condition which was 
actually realized both in Flyracodon cf. H. leidyanus, 
and in Ccenopus cf. G. tridactylus metalophus. This 
partial separation is brought about by a vertical groove 
on the inner side of each tooth which, theoretically, is the 
predecessor of the normal, transverse, median valley. 
There is a small internal basal cingulum cutting across 
the base of this groove on P 4 . 


E. L. Troxell — Horned Eocene Ungulates. 

The protoloph, the anterior cross ridge, on P 34 is 
patterned after the molars; but the metaloph is 
extremely small and irregular. It consists of a small 
tubercle situated between two folds of enamel extending 
inward from the proto- and tritocones, and is distinctly 
separated at the ends from both the ecto- and protolophs 

From the drawings it may be seen that the main 
valleys of the premolars open backward instead of inward 
and in this respect are like those of Ccenopus platy- 
cephalus nanolophus. 

10258 TYPE 
Y. P. M. 

Fig. 5. — Metahyrachyus bicornutus, gen. et sp. nov. Crown view of upper 
teeth and part of skull. The diastema between the canine and molar is 
reduced to half its normal length of 25 mm. by the distortion of the skull. 
The chief distinguishing features are found in the unusual premolars. X%- 

Of the teeth of the cheek series the second molar is 
largest ; it approaches in size that of Hyrachyus princeps 
Marsh. The third molar is relatively small and its outer 
and posterior sides form a fairly smooth curve, with the 
ectoloph extending slightly beyond its edge. On this 
third molar the posterior cingulum is conspicuous ; a 
small overlapping cingulum across the median valley 
corresponds to the small cusp which lies on the base of the 
protocone of M 2 . 

The inner side of the ectoloph of M 3 is nearly straight 
in this specimen. Back of the point of union of the 
metaloph, only, does it curve outward and away from the 
straight line, where it forms a thin, prominent edge. The 
small crista is far back on the paracone and lies low on 
its base. On M 1 it is higher and farther forward. 

On all the molars, the parastyle, antero-exterior corner, 
is rather small and is not set off from the adjacent proto- 
loph by such sharp grooves as one sees in Hyrachyus 

E. L. Troxell — Homed Eocene Ungulates. 37 

The posterior extension of the ectoloph on M 3 , the form 
of the slender premaxillaries and of the nasals, the 
presence of erect canines, the deep posterior nares, all 
remind one of Amynodon erect us nobis, and it seems that 
there must have been a rather near relationship. The 
notches on the anterior ends of the nasals are found also 
in Gcenopus alius nobis. 

The following are important measurements of Metahy- 
racliyus bicornutus: 


Molar-premolar length Ill 

Molar series, length 64.5 

P 4 , width 21.7 

P 4 . lensth 16.5 

M 2 , width 26.3 


Two Eocene specimens are figured and described here 
which are unusual in having on the nasals slight thicken- 
ings, or centers of ossification, which resemble the horn 
supports in later rhinoceroses. One is the type of the 
well known Colonoceras agrestis Marsh and the other 
the type of Metahyrachyus bicornutus, gen. et sp. nov. 
The latter may be distinguished by the caniniform I 3 , 
the large P 1 , the median cross ridge on P 2 , and the odd 
forms of P 3 - 4 with their strong development of the tetar- 
tocone and much reduced metaloph. 

38 Troxell — Hyrachyus and its Subgroups. 

Art. V. — The Genus Hyrachyus and its Subgroups; by 
Edward L. Troxell. (With Plate I.) 

[Contributions from the Othniel Charles Marsh Publication Fund, Reabody 
Museum, Yale University, New Haven, Conn.] 


Definition of the genus. 
Subdivisions of Hyrachyus. 
Hyrachyus agrestis group. 
H. agrestis Leidy. 
H, oairdianus (Marsh). 
Hyrachyus affinis group. 
H. affinis affinis (Marsh). 
H. implicatus Cope. 

H. intermedins, crassidens, and paradoxus Osborn, Scott and Speir. 
H. affinis gracilis, subsp. nov. 
% H. modestus (Leidy). 
Hyrachyus princeps group. 
H. princeps Marsh. 
H. eximius Leidy. 

H. imperialis Osborn, -Scott and Speir. 



As early as the Eocene period in America we find 
representatives of the great group of animals called the 
rhinoceros. With them, and not far removed, we see 
also the ancestors of the horse and of the tapir, and still 
other closely related forms which were early blotted ont, 
and whose line has been discarded from the material 
of the great evolutionary structure. 

Within the great family Rhinocerotidae, there was 
already a differentiation in the Middle Eocene which gave 
rise to later subdivisions: (1) Amynodon, leading to the 
Amynodontidre, extinct with the Middle Oligocene; (2) 
Metahyrachyus nobis, a possible progenitor of the true 
rhinoceros through Trigonias and Ccenopus; and (3) 
Hyrachyus, giving rise to the Hyracodontidse, which 
existed throughout the Oligocene. 

The present study deals with the last of these three 
groups, the cursorial hyracodonts, and with the single 
genus Hyrachyus Leidy. 

As to what constitutes Hyrachyus there need be little 
doubt ; there are about one half dozen valid species which 

Troxell — Hyrachyus and its Subgroups. 39 

may be combined into a harmonious genus with certain 

well defined features and yet with certain individuali- 
ties which indicate a rather wide variation. 

As to what constitutes the type of the genus, there 
has been much misunderstanding. Through a passive 
tolerance, authors have come to accept H. agrarius 
Leidy as the type, following Leidy, who himself chose 
to eliminate H. agrestis, a species founded upon the lower 
jaw of a young animal with milk teeth. 

Although Leidy stated specifically in his first descrip- 
tion of the genus (1871 A, p. 357) that Hyraehgus "an 
extinct genus, allied to Hyraeodon, is founded on a frag- 
ment of a lower jaw of a young animal/' yet in a later 
paper (1873, p. 60) he figures this jaw but says that he 
regards it as the same species as H. agrarhis, and sub- 
stitutes the latter as the genoholotype. This has been 
followed in the literature with only one or two excep- 
tions. It is of no great moment which of these species 
represents the genus, but the rules of nomenclature 
demand that we adhere strictly to the original type in 
spite of the desires of later writers, including the nomen- 
clator himself, and so H. agrarius must give place to 
H. agrestis. 

This species is reinstated as the genoholotype with 
greater confidence because of the following statement 
from Doctor 0. P. Hay, who has recently looked into 
this question and says in a letter dated June 18, 1921, 
"It is evident that Leidy meant to base his genus on 
H. agrestis." 

Neither the holotype of H. agrestis nor that of H. 
agrarius is specifically determinable. The first is based 
on a lower jaw fragment containing the first and fourth 
deciduous premolars, the roots of the intervening two, 
and also the first molar of the permanent series ; the 
second type is founded upon a lower jaw without the 
crowns of any teeth. 

Definition of the Genus. — Hyraeliyus Leidy may be 
distinguished as having smooth nasals without horn 
rugosities ; diastemata separating the canines and 
premolars; canines moderately long, pointed, and only 
slightly flattened; premolars with cross crests not 
parallel but forming a loop, the metaloph encircled by 
the protoloph; molars with strong parastyles consti- 

4:0 Troxell — Hyrachyus and its Subgroups. 

tuting isolated cones, with median valleys open, meta- 
cones receded, with parallel cross ridges leading from 
points anterior to the two main onter cones respectively, 
and finally with a strong posterior extension of the ecto- 
loph on the last npper molar ; dental formula 3. 1. 4. 3. 

All the species of the genus are from the Bridger beds 
of the Middle Eocene. 

The Subdivisions of Hyrachyus. 

Hyrachyus agrestis Group. 

Hyrachyus agrestis Leidy 1871, genoholotype. 
Hyrachyus agrarius Leidy 1871, synonym. 
Hyrachyus bairdianus (Marsh) 1871. 

This group, besides the two species of Leidy already 
discussed, includes also H. bairdianus Marsh, probably 
a subspecies under H. agrestis. 

Leidy 's types, inadequate for accurate specific deter- 
mination, serve only to define the genus. Marsh's type, 
although also far from being complete, offers some dis- 
tinguishing characters and is here redescribed along with 
a nearly complete skull and jaws in the Yale collection 
(Cat. No. 11081, Y. P. M., apotype). 

Hyrachyus bairdianus (Marsh). 

(Figs. 1, 2.) 

Cotypes, Cat. Nos. 11035 and 11057, Y. P. M. Eocene (Bridger), near 
Fort Bridger, Wyoming. 

The first type consists of a portion of the left maxil- 
lary with the three molars (fig. 1), of which M 1 and M 2 
are so worn and broken that one gets little character of 
the species from their study. M 2 shows a strong internal 
cingulum across the transverse valley. The third molar 
is, however, well preserved, and shows certain important 
specific features. The tooth is wide on the inner side, 
narrow on the outer; thus the ectoloph is short, and its 
posterior end, making up the metacone and style, does 
not extend beyond the bulging posterior cingulum. 

The posterior fossette cuts deeply, forming a sharp 
angle between the ectoloph and metaloph. A strong cin- 

Troxell — Hyrachyus and its Subgroups. 


gulum surrounds all of the tooth except the outer side and 
the base of the lrypocone. 

The second specimen (fig. 2) described by Marsh con- 
sists of a portion of the left mandibular ramus with the 

Fig. 1. 

J 103 5 TYPE 
Y. P. M. 

Fig. 1. — Cotype, Hyrachyus bairdianus (Marsh). Part of the maxillary 
with three molars. Nat. size. 

three lower molars. Although the teeth agree closely 
in their width, M 3 is 3 mm. larger than M : . The cross 
ridges of Mj are most nearly parallel, those of M 3 
least so. 

Fig. 2. 

71057 TYPE 
Y P M 
Fig. 2. — Cotype, Hyrachyus bairdianus (Marsh). Crown and side view 
of the lower molars, facing to the left, Xote the uniform size of these teeth. 
Nat. size. 

42 Troxell — Hyrachyus and its Subgroups. 

The new specimen of H. bairdianus, apotype, Cat. No. 
11081, Y. P. M., discovered in the Bridger beds near 
Fort Bridger, Wyoming, by 0. Harger in 1871, consists 
of a fairly complete skull and jaws, with all the dentition 
except I 2 . 3 and F 1 . The series of lower teeth measure 
precisely the same as do those of the type of H. agrarius 
Leidy; the specimen therefore serves as a unit of meas- 
ure for this group so ill denned by the types, and in addi- 
tion gives us an idea of the character of the anterior teeth. 

Hyrachyus affinis Group. 

Hyrachyus affinis (Marsh) 1871. 

Hyrachyus implicatus Cope 1873. 

Hyrachyus intermedins Osborn, Scott and Speir 1878. 

Hyrachyus crassidens Osborn, Scott and Speir 1878. 

Hyrachyus paradoxus Osborn, Scott and Speir 1878. 

Hyrachyus affinis gracilis, subsp. nov. 

f Hyrachyus modestus (Leidy) 1870. 

This group of species, here associated about H. affinis 
(Marsh) as a center, represents the smallest of the 
hyrachyids. Seven names are listed, only four of which 
are thought to be of value in our classification of the 
early rhinoceroses ; H. crassidens, H. paradoxus and 
?H. modestus are based on imperfect specimens, inade- 
quate as types. The three others are here considered 
as subspecies. H. affinis intermedius Osborn, Scott and 
Speir, founded on the upper molars, is slightly smaller 
than //. affinis affinis, especially in the antero-posterior 
dimension, due probably to wear; it is noted for the 
cingulum which nearly encircles the tooth, the obscure 
crista, the cross ridges arched forward, and the position 
of the postero-external lobe far to the rear. H. affinis 
implicatus Cope has for cotypes three specimens, the 
first of which varies in measurements from H. affinis 
affinis in having a shorter last molar but otherwise being 
larger; it has been referred by Cope (1884, p. 675) and 
others to H. agrestis Leidy, while the second cotype is 
presumed to stand for a species under the same name. 
It is the opinion of the writer that the species must stand 
on the merits of its first type, and that, in order to make 
the second specimen a valid type, it must be renamed. 

Troxell — Hyrachyus and its Subgroups. 43 

Hyrachyus affinis affinis (Marsh). 

(Fig. 3.) 
Holotype, Cat. No. 12530, Y. P. M. Eocene (Bridger), near Grizzly 
Buttes, one mile from Marsh's Fork, Wyoming. 

Marsh, in his first description (1871, p. 37), character- 
izes this species by the small size, by the contour of the 
crown of M 3 which has a deep notch in the posterior 

Fig. 3. 

125 30 TYPE Y. P. M. 

Fig. 3. — Holotype, Hyrachyus affinis (Marsh). Crown view of the molars 
and fourth premolar. This is one of the smallest species of the genus. 
Nat. size. 

margin of the base, and by the small, prominent antero- 
external tubercle not closely connected with the adjoining 

In addition to this, one notes the squared form of 
the last molar, the rather weak posterior extension of the 
ectoloph, the large second molar, and the weak metaloph 
on the fourth premolar. 

Hyrachyus affinis gracilis, subsp. nov. 

(PL I and text fig. 4.) 
Holotype, Cat, No. 11170, Y. P. M. Eocene (Bridger), Henry's Fork, 

The type consists of a good skull and jaws, without 
the anterior teeth, fore and hind feet and fore limbs, 
together with parts of the hind limb bones, pelvis, ribs 
and vertebrae. This specimen, previously on exhibition 
for several years in the old Peabody Museum, has just 
been remounted (see pi. I) on a large slab repre- 
senting the original rock matrix; the preparation was 
done by Mr. Hugh Gibb. Professor Lull has made a 
restoration of the animal in the flesh in low relief, one- 
fourth the natural size, modelled after the mounted 

44 Troxell — Hyrachyus and its Subgroups. 

specimen (in pi. I, about one thirteenth nat. size). The 
posture selected for the restorations is that of an 
extreme phase of a fast trot, with alternate front 
and hind feet advanced, the whole purpose being to 
emphasize the idea that this is a light, agile animal, 
a cursorial rhinoceros. The missing parts were restored 
in plaster and the bones arranged on the flat base; the 
position of each was carefully worked out by the study 
of other skeletons, photographs, and observations on 
living animals. 

This interesting specimen, found probably in the early 
seventies, drew the attention of Professor Marsh, who 
apparently intended to make it the type of a new species ; 
his observations follow, but of course the species name, 
here quoted from his manuscript, has no standing : 

"Notes on Hyrachyus bairdi M. 

"No. 1170. Skull and skeleton nearly complete (H. Fork, 

"1. Hind feet longer and stouter than fore feet. (Reverse 
true in tapir.) Metatarsals curved. 

"2. No supratrochlear foramen in humerus (as in tapir). 

"3. Navicular sesamoids present, but minute, in all three toes 
behind, and at least II, III and IV in front. (Not present in 
tapir or rhinoc.) 

' ' 4. Ulna smaller than in tapir. ' ' 

Additional points on the new type. — The dental for- 
mula is 3. 1. 4. 3. P 2 has a strong cross ridge anteriorly 
like the larger premolars; it is slightly wider (8.7 mm.) 
than it is long (8 mm.). P 3 - 4 are very similar in propor- 
tions; the latter is larger. P 1 was probably as wide 
transversely (6.4 mm.) as it was long; the anterior part 
is broken away. The lower premolars show a gradual 
change from the first to the fourth. F ± has a single cusp 
with a small internal ridge leading from the apex; P 2 
shows an antero-internal groove set off by what may be 
called the paraconid; P 3 . 4 both have the two cross 
crests and, the latter especially, take on the form of a true 

The molars increase gradually in length from the first 
to the third, but change very little in width. There is 
practically no difference in height between the anterior 

Troxell — Hyrachyus and its Subgroups. 


and posterior cross crests. In the worn condition these 
crests come to show the typical "L" of the rhinoceros- 
like animals because of the low ridges extending from 
them on the onter forward side. The ridge from the 
anterior crest continues on into an upfolded edge or 

Fig. 4. 

J1170 TYPE 

Y. P. M 

Fig. 4. — Holotype, Hyrachyus affinis gracilis, subsp. nov. Crown views of 
the upper and lower teeth. Note the odd-shaped P 4 and M 3 and the rhinoc- 
eroid lower teeth with the strongly L-shaped cross crests. Nat. size. 

This new subspecies is put under H. affinis (Marsh) 
on account of the agreement in size. Such parts as can 
be compared exhibit some notable differences : the new 
form may be distinguished (1) by the weaker paracone 
on M 3 , which blends into the metacone without such a 
deep groove on the outer side ; (2) by the less prominent 
base of the metacone, giving a more nearly rounded out- 
line to the posterior and outer sides; and (3) by the 
longer P 4 with its heavier metaloph continuous with the 
end of the curved protoloph. 

Hyrachyus princeps Group. 

Hyrachyus eximius Leidy 1871. 

Hyrachyus princeps Marsh 1872. 

Hyrachyus imperialis Osborn, Scott and Speir 1878. 

These species are grouped together because they can 
not be distinguished specifically. Leidy 's type was the 

46 Troxell — Hyrachyus and its Subgroups. 

first studied and named ; it is based on a fragment of the 
mandibular ramus with two teeth, one partly broken 
away and both well worn. The species shown by Leidy 
on his plate IY, figures 19, 20, is clearly inadequate as 
a type ; it is somewhat smaller than H. princeps. 

Hyrachyus imperialis Osborn, Scott and Speir (1878, 
p. 50) was "established on the second and third molars 
of each upper jaw, and three premolars, and one lower 
molar, portions of the skull and vertebrae." The 
measurements agree precisely with the type of H. prin- 
ceps, and so far as one can judge the description also 
tallies. II. imperialis is either a synonym or at most only 
a subspecies of H. princeps. 

Hyrachyus princeps Marsh. 

(Fig. 5.) 

Holotype, Cat. No. 11157, Y. P. M. Eocene (Bridger), Fort Bridger, 

In the first description of this species (1872, p. 125) 
Marsh characterizes it as the "largest of the Tapiridae 
yet found in this country. ' ' He further says : 

"The remains representing it indicate an animal nearly three 
times the bnlk of Lophiodon [Hyrachyus'] Bairdianus Marsh, and 
probably twice that of the individual named Hyrachyus eximius 
by Dr. Leidy. The specimens on which the species is based con- 
sist of a nearly complete series of upper teeth, and several lower 
molars, .... remarkably well preserved. The last two upper 
molars are unusually large in proportion to the rest of the series, 
and have the antero-external lobe quite separate, and with its 
apex incurved. ' ' 

The following additional features will serve to define 
the type of H. princeps more specifically. On the pos- 
terior side of M 3 there is no deep fossa ; the cingulum is 
weak here and is almost lacking on the inner side of the 
tooth, leaving the median valley practically unobstructed. 
The antero-external lobe is heavy and broad, and though 
distinct, is not far removed from the curved protoloph. 
The anterior and posterior sides of the tooth are parallel, 
squaring the tooth to a degree duplicated in H. affinis 
affinis only, within the genus. The straight exterior 
and posterior sides unite to form a prominent angle 

Troxell — Hyrachyus and its Subgroups. 


which makes the posterior end of the ectoloph seem 
small in comparison. The crista, the fold of enamel 
on the inner side of the paracone, is relatively small on 
each molar. 

The first and second molars have forms qnite different 
from the third, due mostly to their larger metacones and 
the greater posterior extension of the ectolophs which 
reach ont beyond the general border of the crown, in a 
strong inwardly curved fold. Just behind the point of 
union of the meta- and ectolophs is the distinct metacone, 
on the outer side of which is a basal ridge or cingulum. 

Fig. 5. 

71157 TYPE 

v. p. m 

Pig. 5. — Holotype, Hyrachyus princeps Marsh. Molars and premolars of 
a very large hyrachyid. M 3 is transposed from the right maxillary. y(2/2>. 

Marsh has mentioned the small size of M 1 ; its diame- 
ters are 26 mm. transverse and 21 mm. fore and aft. 

On the premolars the anterior and posterior sides 
converge inward. They are rounded on the inner side, 
not angular as in the rhinoceroses ; they are squared on 
the broad outer sides, where distinct grooves separate 
the two main cones and the anterior and posterior styles. 

The deuterocone dominates the inner half of the tooth, 
and the protoloph supersedes the metaloph, which is 
small and thin and resembles that of certain species of 
the early rhinoceroses. The inner slope of the deutero- 
cone is not encircled by a cingulum. The parastyle of 
the premolar is distinct but much smaller than in the 

On P 3 the protoloph does not actually join the ectoloph ; 
both it and the anterior cingulum are somewhat broken 
and interrupted. 

P 2 has an incomplete metaloph, as in Ccenopus nanolo- 
plnis, which does not reach the deuterocone, and the 


Troxell — Hyracliyus and its Subgroups. 

protoloph is entirely severed from the ectoloph by the 
deep valley running longitudinally. 

The third lower molar is broad on its posterior half. 
There is a strong basal ridge both fore and aft, the true 
cingulum, in addition to the sharp ridge running inward 
from the antero-exterior corner of the crown. 


The genus Hyracliyus Leidy, taken as a whole, is made 
up of a number of groups, each of which, in the writer's 
opinion, belongs to a single species. H. affinis (Marsh) 
represents the smallest of these, then in the order of size 
come H. bairdianus (Marsh) (=H. agrestis Leidy, 
indet.), and the very large II. princeps Marsh. 

Under the name H. affinis gracilis el new subspecies 
is described which, because of the completeness of the 
skeleton upon which it is based, serves admirably as a 
protype, or reference type, further to establish the group. 
This specimen has recently been remounted. 

Measurements of Types. 

H. bairdianus 









No. 11035 

No. 11.081 

Holotype Holotype Holotype 






Superior dentition : 

Molar -premolar series, 




f 128 

Molar series, length 






P 4 , width, transverse 





P 4 , length 





M 2 , width 





M 2 , length 


16*. 7 



M 3 , width 






M 3 , length 





26 * 

Inferior dentition : 

No. 11057 

Molar-premolar series, 




Molar series, length 




P 4 , width 



P 4 , length 



Mi, width 




M x , length 




M 2 , width 





M 2 , length 




M 3 , width 





Ms, length 





Am. Jour. Sci. Vol. IV, 1922. 

Plate I. — Hyrachyus affinis gracilis, subsp. nov. Above, mounted skele- 
ton. Below, flesh restoration by Professor E. S. Lull. X 1/13. 

Troxell — Hyrachyus and its Subgroups. 49 


Cope, E. D. 1873. On some Eocene mammals obtained by Hayden's Geo- 
logical Survey of 1872. Pal. Bull. No. 12. 

— 1881. The Vertebrata of the Tertiary formations of the West. Eept. 
U. S. Geol. Survey Terr., 3, 657-77. 

Leidy, Joseph. 1870. [Remarks on a collection of fossils from the western 
territories.] Proc. Acad. Nat. Sei. Philadelphia, 22, 109. 

— 18 71 A. Eeport on the vertebrate fossils of the Tertiary formations of 
the West. U. S. Geol. Survey Wyoming and portions of contiguous 
Territories, 2d (4th) Ann. Eept. 

— 1871B. Remarks on fossil vertebrates from Wyoming. Proc. Acad. Nat. 
Sci. Philadelphia, 23, 229. ' 

— 1873. Contributions to the extinct vertebrate fauna of the western terri- 
tories. Rept. U. S. Geol. Survey Terr., 1. 

Marsh, 0. C, 1871. Notice of some new fossil mammals from the Tertiary 
formation. This Journal (3), 2, 35-44. 

— 1872. Preliminary description of new Tertiary mammals. Ibid. (3), 4, 

Osborn, H. F., Scott, W. B., and Speir, Francis, Jr. 1878. Palseontologieal 
report of the Princeton scientific expedition of 1877. Contrib. Mus. 
Geol. and Arch. Princeton Coll., 1. 

Am. Jour, Sci.— Fifth Series, Vol. IV, No. 19. — July. 1922. 

50 C. W. Cook — New Occurrence of Ilsemannite. 

Aet. VI. — A New Occurrence of Ilsemannite; by Chas. 

W. Cook. 

The molybdenum-bearing mineral ilsemannite seems 
to be of sufficiently rare occurrence to make such occur- 
rence worthy of note. It has been reported from Blei- 
berg, Carinthia ; x Cripple Creek, Colorado ; 2 Natal, S. 
Africa f Ouray, Utah; 4 and as possibly occurring in the 
mine waters at Idaho Springs, Colorado. 5 During an 
extensive investigation of molybdenum deposits, in which 
a large number of localities in the United States and 
Canada were visited and hundreds of specimens from 
localities not visited were examined, the writer observed 
the occurrence of ilsemannite at but two points ; one at 
Ouray, Utah, mentioned above, and the other in Shasta 
Co., California, four miles west of Gibson, a station on 
the Southern Pacific R. R. 

In the Gibson locality molybdenite, associated with 
pyrite, occurs disseminated in what appear to be bowlders 
of aplite, no definite connection between the molybdenite- 
bearing masses and other rocks having been established. 
Dikes of similar lithological character were found in the 
district but in no instance, where it was certain that the 
rocks were in place, was any molybdenite found. Not 
infrequently the molybdenite has been completely altered 
to molybdite and this mineral may be seen entirely filling 
the cavities formerly occupied by the molybdenite. In 
other instances, where the molybdenite has not been com- 
pletely altered, a bluish-colored zone shows around the 
molybdenite nucleus. This zone is plainly a staining of 
the quartz and feldspar by ilsemannite or some closely 
allied substance. 

At the time of the examination of the property, some- 
thing over one hundred tons of ore had been removed, 
according to reports, and some of this ore was still on 
the ground at the railroad awaiting shipment. This 
broken rock had been exposed to the action of the atmos- 
phere for several months and many pieces had been com- 
pletely discolored blue on the surface, giving them an 
entirely different appearance from that of the freshly 

1 Hof er, N. : Jahrb. Min., 1871, p. 566. 

2 Lindgren and Eansome : U. S. G. S. Prof. Paper No. 54, p. 154, 1906. 
3 duToit: So. African Jour. Sci., vol. 13, p. 153, 1917. 

4 Hess: U. S. G. S., Min. Kes., 1917, p. 913, 1920. 
B Horton: U. S. Bur. Min., Bull. Ill, p. 15, 1916. 

C. W. Cook — New Occurrence of Ilsemannite. 51 

broken rock. It would seem, therefore, that the altera- 
tion of the molybdenite to ilsemannite had been relatively 

According to different authorities, ilsemannite has 
been formed in different ways. Cahen and Wooten' 5 state 
that it is formed by the alteration of jordisite ; Dana 7 says 
that ilsemannite is "a product of the decomposition of 
metallic molybdates"; while Lindgren and Ransome 8 
indicate their belief that at Cripple Creek it has been 
formed by the direct oxidation of molybdenite. 

The exact composition of ilsemannite is likewise sub- 
ject to discussion. Dana 9 gives the formula (Mo0 2 . 
4Mo0 3 ) ; Schaller 10 has proposed the formula (Mo0 3 . 
S0 3 .3H 2 Q) ; while Yancey 11 believes that Guichard's 
formula for the synthetic blue oxide (Mo0 2 .4Mo0 3 .6H 2 G) 
most closely, although not exactly, expresses the compo- 
sition of the blue of ilsemannite and for the present 
prefers to consider it is a chemical mixture of molyb- 
denum dioxide with relatively larger amounts of the tri- 
oxide. In support of the latter supposition, the writer 
wishes to advance some observational evidence which 
may also shed some light on the nature of the alteration 
process in the case of molybdenite. 

On the crystal and cleavage faces of the molybdenite in 
specimens from Gibson, a phenomenon has been observed 
which the writer has not seen on specimens from any 
other locality, namely a tarnish or iridescence which 
immediately reminds one of the "peacock colors" on 
bornite. A closer examination shows the color to vary 
from a bronze-brown to a violet brown to blue. Appar- 
ently the product is identical with that obtained near 
the assay on charcoal when molybdenite is subjected to 
the action of the oxidizing flame. Also the writer has 
produced this tarnish artificially by gently touching 
crystals and cleavage faces of molybdenite with an oxi- 
dizing flame. Guichard 12 has shown that the true color 
of molybdenum dioxide is brown or violet brown although 
this color is easily obscured by the blue of the mixed 
oxides. ^ It, therefore, seems to the writer quite probable 
that this bronze -to violet brown coating on the molyb- 

c The Mineralogy of the Bare Metals. 1912, p. 51. 

7 System of Mineralogy. 6th edition, p. 202. 

8 Loc. cit. 

9 Loc. cit. 

10 Wash. Acad. Sci., vol. 7, p. 417, 1917. 

11 Chem. and Met. Engr., vol. 19, p. 189, 1918. 

12 Compt. Eend., vol. 129, p. 722, 1899. 

52 C. W. Cook — New Occurrence of Ilsemannite. 

denite from Gibson is molybdenum dioxide. All material 
so far examined indicates a thickness for the coating no 
greater than that of one of the cleavage laminae of the 
molybdenite. Hence, the definite determination of its 
composition or of its properties other than color and 
luster, the latter being metallic, has been impossible. 

One other observed phenomenon should- also be men- 
tioned since it may have a direct bearing not only upon 
the composition of both the coating and ilsemannite but 
also upon the nature of the decomposition process. In 
some instances the blue colored material which has been 
designated as ilsemannite does not appear to be water 
soluble. This might be expected if the following stages 
were passed through in the alteration of molybdenite to 

If the molybdenite was first altered by the oxidation 
of the sulphur, molybdenum dioxide, a brown substance 
insoluble in water (Muthmann), 13 would result. This 
would represent the tarnish stage mentioned above. The 
subsequent oxidation of a small portion of the molyb- 
denum to the hexavalent form would result in a change 
of color from brown to blue although the compound would 
still remain insoluble in water, and would correspond to 
the substance referred to in the preceding paragraph. 
The existence of such an insoluble blue compound might 
be expected from the fact that for many years after its 
original preparation, the color of molybdenum dioxide 
was thought to be blue. Further oxidation, with an 
increase in the relative amount of the trioxide, and hydra- 
tion would then transform this insoluble blue compound 
into the soluble blue compound, ilsemannite. Finally, 
complete oxidation to the trioxide, on combination with 
iron, would yield the yellow compound, molybdite. 

The above interpretation of the observed facts would 
seem to indicate the composition proposed for the coating 
and also to support Yancey's suggestion regarding the 
composition of ilsemannite. Further, it has a direct 
bearing upon the question of the secondary enrichment 
of molybdenite, a subject now under investigation by the 

University of Michigan, 
Ann Arbor, Mich., 
February 14, 1922. 

13 The process of Ullik described by Muthmann (Liebigs Annalen vol. 238, 
p. 114, 1887) involves purification with hydrochloric acid and potassium 
hydroxide so that it seems probable that it is insoluble in water. 

Perner & Kodym — Silurian of Bohemia. 

Art. VII. — On the Zonal Division and Correlation of the 
Silurian of Bohemia; by J. Perxer, with the collabora- 
tion of 0. Kodym. 

The Silurian formation in central Bohemia (= Upper 
Silurian in the sense of European geologists) is repre- 
sented by three "bandes," designated by Barrande as 
E 1? E 2 , and F t . The boundaries of these beds were orig- 
inally not sufficiently defined, and especially that between 
the most important beds, Ej and E 2 , which was somewhat 
arbitrary owing to the gradual passage of one into the 
other ; there has also been a change of opinion as to where 
and how the boundary line should be drawn between Ej 
and E 2 . No attempts have been made to divide all these 
beds, formed chiefly of very fossiliferous shales and lime- 
stones and attaining at several localities a thickness of at 
least 500 feet, into minor divisions or zones, as has been 
very successfully done in other countries. However, Marr 1 
and Tullberg, 2 in discussing Barrande 's theory of the 
so-called "colonies," have ascertained the existence of 
several graptolitic zones in E x similar or analogous to 
those in England and Sweden. "Wenzel, 3 on the contrary, 
comparing the Lower Paleozoic deposits of Bohemia with 
those of Great Britain, endeavored to prove that the 
geological distribution of graptolites in Bohemia is quite 
different, and quoted associations of graptolites in E t 
(on the same slab) which in other countries appear sep- 
arately in altogether different horizons, and declared that 
the fauna in the Silurian basin of Bohemia was so con- 
centrated that no such zones could be distinguished. 

In order to get a reliable basis for a more detailed divi- 
sion of our Silurian and a closer comparison with that in 
other countries, I undertook a revision of the Bohemian 
graptolites, 4 which had been neglected from the paleonto- 
logical standpoint since 1850. 5 I found in many cases that 
Barrande had included more than one species under a 
single specific name, and in some instances had even 

1 Quart. Jour. Geol. Soe., London, 1880. 

2 Sveriges geolog. Undersokning, Ser. C, Xo. 50, 1882; Zeits. deut. geol. 
Gesell, 1883, 2. 

3 Jahrb. d. geol. Reiehsanstalt, Wien, 41, 1, 1891. 

4 Etudes sur les graptolites de Boheme, I, II, Ilia, b, with 17 plates. 
Prague, 1891-1899. 

5 J. Barrande, Graptolites de Boheme, with 4 plates, Prague, 1850; 
E. Suess, Leber bohmische Graptoliten. Haidiger's naturwiss. Abhandl., 
11, 1851. 

51 Perner & Kodym — -Zonal Division and 

included three species confined to different, vertically 
widely separated horizons. Consequently correct identifi- 
cations could not be made until such a revision had been 
finished, and this explains not only the quite erroneous 
assertions of Wenzel as to the non-existence of zones in 
the Silurian of Bohemia, but also the incongruous views 
on graptolitic zones in Bohemia held by other authors. 
In addition, a more detailed division and comparison of 
Bohemian strata could not be systematically worked out 
until this revision of species had been made. After 
redefining the world-widely distributed species of Bar- 
rande in accordance with his type specimens, I was able 
to ascertain in the Silurian beds E 1? E 2 , and F x many new 
species, and to note the presence of forty forms heretofore 
known only from Great Britain and Scandinavia. Addi- 
tional studies regarding the vertical distribution of other 
fossils and different facies in the beds named above have 
led to results here communicated. 6 

The limit between the Upper Ordovician (Barrande's 
Bande D 5 ) and the Lowest Silurian beds in central 
Bohemia is a very sharp one, both from the petrographic 
and the paleontologic standpoint. The lowest Silurian 
beds, black graptolitic shales of E 1? rest at some rare local- 
ities with a slight unconformity on the soft yellow or 
olive-green, sometimes argillaceous, shales of D 5 , which 
are interstratified, especially toward the top, with quartz- 
ites, gritty shales, and graywackes. The fauna of D 5 is 
also completely different from that in E, having no species 
in common, and being characterized by Trinucleus, Remo- 
pleurides, Carmon, Dindymene, Homalonotus, 2Eglina, 
Asaplius, Agnostus, Areia, and Dicellograptus, as against 
the Ej shales bearing Diplograptus, Rastrites, and Mono- 
graptus, so that a hiatus seems to be evident between the 
Ordovician and Silurian in Bohemia. 

6 It was my intention immediately after the completion of my paleontolog- 
ical studies on Bohemian graptolites to publish as their final part a paper 
dealing with the zonal division of the graptolitic rocks of Bohemia. But 
my official duties connected with the Bohemian Museum and with working out 
the Gastropoda of Barrande prevented my doing this before the outbreak of 
the Great War. During the past few years I have been greatly aided by my 
pupil, Mr. Od. Kodym, with whose collaboration a preliminary report on this 
subject was published in the Journal of the Bohemian Museum for 1919 (in the 
Czech language)' In the present English communication I hope to enable 
foreign workers to get the review and comparison of the Bohemian Silurian 
beds before the more detailed paper on this subject appears, since I have 
no idea when this latter can be accomplished, because of the desolate situa- 
tion of the scientific institutions of Czekoslovakia. 

Correlation of the Silurian of Bohemia. 55 

Bande E x is composed chiefly of black graptolitic shales, 
but becomes calcareous toward the top, passing into the 
limestones of E 2 . This passage is marked by the appear- 
ance of calcareous concretions (anthracolites) and thin 
seams of bituminous limestones in the higher graptolitif- 
erous shales. These intercalations become more numerous 
toward the top of E 1? until the shales gradually dis- 
appear and limestone-beds prevail, followed by thickly 
bedded compact or crystalline limestones. The boundary 
between Ej and E 2 , formerly uncertain, was proposed by 
J. Jahn 7 almost wholly on the basis of petrography, and 
the horizon of shales with concretions and limestone inter- 
calations was classed as E 2 . I myself defined the same 
boundary paleontologically, s taking the zone with Mono- 
graptus colonus, M. dubius and M. roemeri as the highest 
horizon of E^ This limit agrees in many localities with 
the petrographic one of Jahn, and I compiled at the same 
time a list of 230 more common molluscs, brachiopods, 
and crustaceans, indicating their true horizon, E x or E 2 , 
based on the new material collected from the chief expo- 
sures zone by zone ; this will, I hope, aid in correcting the 
designations in many collections abroad (see pp. 64-66). 

Bande E x . 

E 1? as now limited, can be divided into three chief subdi- 
visions, which in general correspond to three "zones" 
recognized by Marr, and for which I propose accordingly 
(as in use for other Lower Paleozoic rocks of Bohemia) 
the following designations : 

E x a = Diplograptus beds. 
E x /3 = Priodon beds. 
Ej-y = Dubius beds. 

About thirty years ago Jahn (1. c.) proposed to desig- 
nate as E^ the graptolitic shales in which the calcareous 
concretions are absent, and the upper part of Ej (shales 
with concretions and limestone intercalations) as Ej/?. I 
could not accept this purely lithological division, as it is 
inaccurate and unreliable. As the far more reliable 
graptolite fauna demonstrates in many sections, there are 
shales containing Monograptus priodon, M. vomerinus, 
Stomatograptus grandis, Cyrtograptus murchisoni, and 

7 Jahrb. d. geolog. Reichsanstalt, Wien, 42, 3, 1892. 

8 Bohemian Acad. Sci., Jubileum Memoirs, XX, 1915. 

56 Perner & Kodym — Zonal Division and 

Retiolites geinitzianus in some localities without nodules 
and limestone intercalations; in others, however, they 
are crowded with them, although they both correspond to 
Marr's "priodon zone." The same thing occurs in some 
upper zones. The division into three parts is, I am con- 
vinced, more appropriate to the three distinct graptolitic 
faunas existing in Ej and in general fits to the facts known 
from other countries. 

E x a — Diplograptus Beds. 

Eja or the Diplograptus beds is characterized by the 
frequent occurrence of many species of Diplograptus and 
other genera of the Diprionidse, which are all absent in 
the upper beds. It contains four zones, as follows : 

a. Zone of Diplograptus vesiculosus Nich. — Soft, 
light yellow or brown, lilac-shaded shales, known only from 
Belec* and Libomysl. Aside from D. vesiculosus and an 
undeterminable Climacograptus , no other graptolites are 
known from this zone, demonstrating that in Bohemia, 
beds of Lower Llandovery age are also represented. In 
the other localities this zone can not be ascertained, as for 
instance, in the vicinity of Tman, where similar shales 
resting directly on D 5 (Ordovician) have so far yielded 
no graptolites. At Korno and elsewhere black shales 
lying on D 5 are metamorphosed by diabase intru- 
sions into hard hornstone-like shales without a trace of 
graptolites. At several localities this zone is certainly 
absent (for instance, at Zadni Treban), because the black 
shales lying directly on D 5 belong already to the next 
higher zone (Rastrites peregrinus). It seems that the 
transgression which began after the close of Ordovician 
time gradually reached different places in the ' ' Silurian 
basin" of Barrande. 

b. Zone of Rastrites peregrinus Barr. — This zone also 
contains shales but of different petrographical characters, 
and is developed in the entire area of the Silurian. For 
example, at Treban, Motoly,and colony "Haidinger" 9 
between Radotin and Kuchelbad, the zone consists chiefly 
of soft micaceous, somewhat arenaceous, black or gray 
shales ; in the vicinity of Tman, of yellow or grayish, lilac- 

* In this and other cases beyond, the accent is from necessity omitted. 
9 The names of Barrande ? s colonies are used here only as designations of 

Cor relation of the Silurian of Bohemia. 57 

shaded, non-micaceous, fine shales ; in most localities, 
however, are found hard, gritty, hornstone-like, black 
shales, especially in the neighborhood of diabases (Karlik, 

Faunistically, this zone is the richest one, and it is 
relatively thick enough to be divided into subzones. 
Besides Rastrites peregrinus, which is everywhere the 
most common species, the following forms occur : Diplo- 
graptus modestus Lapw., D. (Glyptograptus) sinuatus 
Nich., D. (G.) bellulus Tornq., Climacograptus scalaris 
Lin. (auct.), Cephalograptus folium His., C. cometa Gein., 
Retiolites (Gladiograptus) perlatus Nich., Monograptus 
lobiferas M'Coy, M. communis Lapw., M. fimbriatus 
Nich., M. leptotheca Lapw., M. convolutus His., M. tri- 
angulatus Harkn., M. distans Port!., and Corynoides sp. 

None of these species appears in the next higher zone. 

c. Zone of Rastrites linnei Barr. — Very fine-grained, 
fissile shales, almost without mica, mostly of gray-green- 
ish color ; the graptolites in them have a bright silver-like 
luster ; the best locality is at Zelkovice, where occur : 
Diplograptus palmeus Barr., D. ovatus Barr., Monograp- 
tus beclil Barr., M. clingani Carr. var. hopkinsoni Per., 
M. halli Barr., M. holmi Per., M. marri Per., M. jaculum 
Lapw. var. variabilis Per., M. planus Barr., M. proteus 
Barr., M. runcinatus Lapw., M. turriculatus Barr. var. 
minor, and Rastrites linnei Barr. 

d: Zone with Monograptus turriculatus Barr. — Sim- 
ilar hard dark shales, in some localities clayish and of 
dark brown color. Characteristic graptolites are: 
Monograptus palmeus Barr. var. tenuis Barr., M. turri- 
culatus Barr. (typical large form), M. jacidum Lapw., and 
Retiolites (Plegmatograptus) obesus Lapw. Some grap- 
tolites of the preceding zone occur sporadically here also, 
as M. proteus, M. runcinatus and M. marri. 

E^fi — P Hod on Beds. 

The commonest fossils in these beds are Monograpti of 
the group of M. priodon Bronn; for the first time appear 
here the genera Cyrtograptus and Stomatograptus ; 
totally absent are Rastrites, Diplograptus, and Climaco- 
graptus. In this horizon we find already in the grap- 
tolitic shales calcareous concretions (nodules of ellipsoidal 

58 Pemer & Kodym — Zonal Division and 

shape, 2-15 inches in diameter), often forming almost con- 
tinuous beds ; thin beds of shaly limestone and compact 
or crystalline limestone often alternating with shales are 
also a common feature in many localities. Some- 
times these calcareous intercalations contain trilobites, 
brachiopods, various molluscs, etc., and graptolites 
in them are often preserved "en relief." As already 
suggested, these intercalations do not occupy a definite 
horizon ; in some sections they are not to be met with until 
the much higher beds. Some of these limestones are 
very similar to those from Barrande's Bande E 2 , and 
many fossils which are confined to these E x limestones 
have indeed been erroneously cited as occurring in E 2 . 
Three zones can be distinguished in E^: 

a. Zone with Monograptus spiralis Gein. var. subconi- 
cus Tornq. — Dark brown fissile shales, or grayish and 
dark flagstones ; no calcareous nodules present. Char- 
acteristic graptolites are: Monograptus priodon Brn., 
M. vomer inus Nich., M. spiralis Gein. var. subconicus 
Tornq., M. griestonensis Lapw., M. sacculiferus n. sp., 10 
and Retiolites geinitzianus Barr. 

b. Zone with Cyrtograptus murchisoni Carr. — Shales 
of similar lithological character to those in the former 
zone. First appearance (in the same localities) of cal- 
careous nodules. In the western part of Barrande's 
"Silurian basin" 11 these shales alternate with dark com- 
pact limestones or light crystalline limestone ; the latter 
occurrence seems to be in connection with the nearby coral 
reef, forming an almost continuous seam between Tach- 
lovice and Tetin, and reaching vertically into higher zones 
(see section on the coral-reef facies). The shales of this 
zone yielded: Monograptus priodon Brn., M. vomerinus 
Nich., M. latus M'Coy, M. kettneri n. sp., M. remotus 
Tornq., Cyrtograptus murchisoni Carr., M. purkynei 
n. sp., M. centrifugus n. sp., M. insectus n. sp., M. evolvens 
n. sp., Retiolites geinitzianus Barr., and Stomatograptus 
grandis (Suess). In the limestone and calcareous shales 
and mudstones at Lodenice I have found, in association 

10 In recent years many new graptolites have been found in E 1? descriptions 
of which are being prepared by my pupil and collaborator, .0. Kodym. 

11 E. Kettner has proposed and is using the term ' ' Barrandien ' ' for the 
area in Bohemia in which are developed the Paleozoic rocks included in 
Barrande 's old and inappropriate term, ' ' bassin Silurien du centre de la 
Boheme, " viz., his Etages A-H (Algoukian to Devonian). 

Correlation of the Silurian of Bohemia. 59 

with J/, priodori, M. vomerinus, Cyrtograptus murchisoni, 
and Betiolites geinitzianus, the trilobites Arethusina kon- 
incki Barr., Acidaspis mira Barr., A. prevosti Barr., 
A. roemeri Barr., and Bronteus planus Corda, which have 
been many times qnoted as E 2 fossils, although they are 
confined to E 2 . With them are associated a few brachio- 
pods, as Leptccna transversalis and Atrypa sappho, which 
are also met with in E 2 ; and several cephalopods extend- 
ing into higher zones occur here. 

c. Zone with Monograptus riccartonensis Lapiv. — 
Chiefly brown or dark fissile calcareous shales with 
nodules and intercalations of dark compact limestones. 
Best localities: Lodenice, Vyskocilka, colony "Krejci," 
Koneprusy. Here have been found: M. riccartonensis 
Lapw., 12 M. vomerinus Nich., M. dubius Suess var. 
proceva, M. solitaries n. sp., 71/. validus Per., and M. prio- 
don Brn. Associated are : Slava bohemica Barr., Apty- 
chopsis prima Barr., and Arethusina koninchi Barr. 

E^/— Dubius Beds. 

These are characterized by the occurrence of Mono- 
grapti of the type of M. dubius Suess, and allied forms, 
such as M. colonics Barr., etc. 13 The shales of these beds 
very often alternate with seams of calcareous nodules and 
limestones, in which, besides the rarer graptolites, some- 
times occur brachiopods, various molluscs, and crinoidal 
remains. In the northwestern region of Barrande's 
" Silurian basin" these beds are developed as coral-reef 
limestones (see section beyond). 

Two zones can be distinguished : 

a. Zone with Monograptus testis Barrande. — Brown 
or grayish yellow, argillaceous, slightly calcareous shales 
without concretions or nodules (Barek), or dark shales 
with concretions and limestone intercalations (Dvorce, 

12 Most of the specimens of this species I formerly considered to belong to 
M. flemingi Salter and allied forms. The admirable monographs of British 
graptolites by G. E. Elles and E. M. E. Wood (Mrs. Shakespear) in Paleon- 
togr. Soc. London 1901-1913 now enable me to undertake a revision of all 
the British species formerly recorded in my ' ' Etudes, ' ' and to correct some 
previous determinations. 

13 The original designation, Marr's " colonus zone", I think should be 
changed to "dubius beds" since M. dubius occurs in the whole of E^, 
while M. colonus is absent in the lower part. 

60 Perner & Kodym — Zonal Division and 

Kozel). This zone yielded: Monograptus testis Barr., 
M. bohemicus Barr., M. flemingi Salt., M. nilssoni Barr., 
M. dubius Suess, Cyrtograptus lundgreni Tullb., and 
Retiolites (Gotliograptus) nassa Holm. In addition have 
been f onnd associated : Ampyx rouaulti Barr., Palman- 
ites orba Barr., Proetus (Phcetonellus) dentatulus Novak, 
Piscina unguis Barr., P. nana Barr., Orthis honorata 
Barr., and Orthotheca fragilis Nov. 

b. Zone with Monograptus colonus Barrande. — Lith- 
ological character nearly the same, brown, somewhat 
argillaceous shales prevailing, concretions frequent; the 
alternating limestones here more frequent, dark, non- 
crystalline, compact (in contrast to the majority of the 
very similar E 2 limestones), sometimes with numerous 
cephalopods or crinoid fragments. Associated occur 
here : Monograptus colonus Barr., M. nilssoni Barr., M. 
bohemicus Barr., M. roemeri Barr., M. goilandicus Per., 
and M. unguiferus Per. In the same zone, the concretions 
and limestones yield, sometimes in association with the 
graptolites named above: Cyrtoceras vestitum, Gompho- 
ceras clava, Ophidioceras proximum, Orthoceras epulans, 
0. socium, 0. illudens, 0. pleurotomum, 0. styloideum, 
Ascoceras cf. murchisoni, Cirropsis spp., Cyrtolites 
eximius, Euryzone tuboides, Orthonychia ampla, 0. ele- 
gans, Platyceras longipes, Polytropis (Poleumita) potens, 
Spirina paiida, S. tubicina, Avicula glabra, Slava 
bohemica, Cromus beaumonti, Cyphaspis, and Plcenus 
bouchardi. The shales of Ejy also sometimes contain 
Orthoceras, some Acephala, and brachiopods, but these 
are mostly undeterminable. 

Reef-facies in Barrande ? s Etage E. 

The upper Ej beds in the northwestern Silurian region 
are developed as coral-reef fades, with coarse crystalline 
limestones of gray or brown color (never white as in F 2 ), 
sometimes indistinctly stratified, or nodular. The lowest 
part of these limestones must be attributed to E 2 beds, as 
their base is formed by the zone of Cyrtograptus murchi- 
soni (Ej/?). They extend up from E x y to the highest E 2 
beds, and it is most probable that they existed in some 
localities without a noteworthy interruption until the mid- 
dle part of Barrande's Bande G u which is of Lower 
Devonian age. 

Correlation of the Silurian of Bohemia. 61 

A boundary between E x and E 2 can not easily be drawn 
in these limestones, as their lithological character is the 
same through the whole thickness of these deposits. 
Graptolites are almost absent and the fauna bearing a 
decided coral-reef character is for the greater part 
unknown from the other "normal" parts of Ei or E 2 . 
From the rich characteristic coral fauna, which occurs 
chiefly at Kozel and Tachlovice, may be named : Holy sites 
catenularia, Farosites asper, F. taclilovicensis, Cyatlw- 
phyllum prosperum, Cystiphyllum bohemicum, and Om- 
phyma grancle. Besides these corals, there frequently 
occur in these deposits: Orthoceras caduceum, 0. endy- 
mion, 0. pedum, 0. valens, 0. cur reus, Modiolopsis pupa, 
Spanila cardiopsis, Cardiola persignata, Conocardium 
dorsatum, Atrypa obolina, A. inelegans, A. reticularis, 
A. squamma, A. dormitzeri, Bhynchonella niobe, R. 
nympha, and R. princeps. The greater part of the Mol- 
lusca and Brachiopoda mentioned above are apparently 
confined to these coral limestones. 14 

In the immediate neighborhood of these coral reefs 
there seems to be a peculiar trilobite fauna, represented 
by Bronteus planus, Clieirurus insignis, Splice rexoclius 
minis, Deiphon forbesi, Liclias, etc., the true horizon of 
which belongs to E 1? as I have found them in the mud- 
stones and limestones associated with Retiolites geinitzi- 
anus, Cyrtograptus murcliisoni, Monograptus priodon, 
and M. vomer inns. 

Baxde E 2 . 

As already stated, the boundary between E x and E 2 was 
proposed by me on a paleontological basis, taking the 
horizon characterized by Monograptus dubias, M. colonus 
and M. roemeri as the highest in E x . As in E x , there are 
to be distinguished in E 2 two different developments of 
deposits or facies : (a) normal, stratified limestones with 
intercalated shales and mudstones; (b) reef facies, with 
nodular, indistinctly or irregularly stratified coral lime- 
stones, a direct continuation of the reef facies mentioned 
above in E T , and not divisible into zones, the faunal char- 
acter remaining the same. 

14 Atrypa reticularis, Rhynchonella nymplia, and F. princeps occur also in 
the coral limestones of the Devonian stage F 2 and extend to G,. 

62 Perner & Kodym — Zonal Division and 

In the normal E 2 there are three chief horizons which 
are recognizable in the entire Silurian area, as follows : 

1. Cephalopod horizon. 15 

2. Brachiopod horizon. 

3. Crinoid horizon. 

Between them can be distinguished several subordinate 
zones, which in one locality, where the exposure can be 
described in detail, can be characterized by their fauna ; 
but in another, not far distant, can not be demonstrated. 
A majority of the E 2 fossils occur in all the E 2 beds ; 
some species, however, very characteristic for one limited 
locality, are totally unknown in others. A detailed 
scheme of sequence applicable to the whole horizontal 
extent of E 2 can not be given at present, but I hope future 
detailed comparative studies of numerous exposures, like 
those of Kiaer made in Norway's Silurian, will throw 
more light on this subject. 

Cephalopod Limestones, 

These are usually grayish or bluish, compact limestones, 
in thick beds, that in some localities have intercalations 
of brown shales and shaly limestones. Abundant species 
are : Orthoceras nobile, 0. bohemicum, 0. janus, 0. explana- 
tum, Cyrtoceras murchisoni, C. vemum, Gomphoceras 
cylindricum, and Liluites (Ophidioceras) simplex. This 
horizon is also the chief source of other well known fossils 
at Dlouha Hora and Hopanina, among them being Cromus 
beaumonti, Calymene baylei, Harpes ungula, Cyphaspis 
burmeisteri, C. depressa, Cheirurus insignis, Proetus 
striatus, Lytospira subuloidea, Murchisonia latona, Lox- 
onema beraunense, Cardiola bohemica, C. interrupta, C. 
grandis, Dualina excisa, and Atrypa sappho. A charac- 
teristic graptolite is the rare Monograptus transgrediens 

Brachiopod Horizon, 

White-gray limestones without shaly intercalations 
(Dlouha Hora), or dark, thinly bedded limestones alter- 

15 These designations do not mean that, for instance, cephalopods in E 2 are 
confined to the lowest beds; they are intended only to indicate the sudden 
and striking appearance of the three groups of animals, so that the sequence 
named above would mean rather a sequence of different bionomic conditions. 

Correlation of the Silurian of Bohemia. 63 

nating with shales (Lochkov). The majority of the 
limestones are finely crystallized. In certain localities, 
some of the limestone beds consist almost exclusively of 
the shells of Atrypa linguata, in other horizons this 
species, is rare. Very common are Pentamerus, Cyrtia, 
Merista, and Rhynchonella. Graptolites are represented 
by Monograptus ultimus Per., which, however, continues 
up to the highest E 2 beds. Records of the occurrence of 
M. priodon or M. colonics in E 2 are quite erroneous. 

Crinoid Horizon. 

Dark bituminous limestones in thick and thin beds with 
numerous crinoid remains (chiefly stem fragments, often 
even loboliths or Camarocrinus), forming locally true 
crinoid limestones. 16 They are followed by brown or 
gray, sometimes nodular limestones, alternating with 
brown shales or shaly limestones, with flat calcareous con- 
cretions. In this horizon occurs Monograptus ultimus 
Per., associated with Cardiola interrupta, cephalopods, 
dendroids (Dictyonema, C alio gr aphis, Besmogr aphis, 
Rhodonograptus), and Gigantostraca, Some of the 
limestone beds yield gigantic Orthocerata, as 0. neptuni- 
cmn, 0. severum, 0. pot ens, 0. pelagicum, 0. temper ans, 
0. extenuatum, 0. socium, and 0. rivale. In general, the 
E 2 beds with their frequent alternations of limestones with 
shales and mudstones, and their changing fauna seem to 
suggest that they are deposits of a shallow sea-basin, not 
distant from islands or coral reefs, where the bionomic 
conditions changed rapidly. 

Distribution of Fossils in E x and E 2 . 

As already mentioned, many fossils have been described 
by Barrande and others as occurring in E 2 although they 
are confined to Ej or at least are common to both levels ; 
this is true chiefly of Mollusca and Crustacea in the lime- 
stones. These errors have been due not only to the uncer- 

16 These limestones are often very similar to those in E^/ (zone with 
Monograptus colonus), which also bear in many localities numerous crinoid 
fragments (stems, loboliths, rarely cups). 

64 Pemer & Kodym — Zonal Division and 

tainty in delimitation of E 1? but also to the similar aspect 
of the limestones occurring both in E^ and in E 2 . 17 

After collecting at the chief localities with due regard 
to the present boundaries between Ej_ and E 2 , and com- 
paring with Barrande's material, I was able to compile 
the following list of the more important fossils found in 
E x and E 2 , indicating their true range. Graptolites which 
have been treated more in detail in describing the E 2 
zones are omitted here, but special attention is given to 
other fossils which have not been known at all from E x or 
not with certainty, and which occur in the limestones. 
The fauna of E l7 as it is now delimited, therefore appears 
much richer than has hitherto been supposed. 

Cephalopoda: Cyvtocevas abstinens CE^ 18 , f ormidandum (1-2) , 
I epidum (1-2), nigrum (1-2), phillipsi (1) , pluios (1-2), potens(l), 
serration (1), sup erbum (1-2), vestitum(l), victor (1) ; Gompho- 
ceras ag as sizi (1-2), amygdala (1-2), at rophum (1-2), billingsi 
(1-2), bohemicum (1-2), clava (1-2), contr avium (1-2) , Jialli (1-2), 
imp eviale (1-2), mumia (1-2), nucifovme (1-2), ovum (1-2), vigi- 
dum (1-2), sphrnvo soma (1-2), vespa (1-2) ; Ophidiocevas amissus 
(1), proximus (1) , rudens (1-2), tener (1-2) ; Orthoceras cequabile 
(1), annulatum Sow. (1-2), b ohemicum (1-2) , compulsum (1-2), 
cuvvens (1-2), decuvtatum (1-2) , docens (1-2), duponti (1-2), egve- 
gium (1-2), endymion (1-2), epidans(l), extensum(l), gevmanum 
(1-2), illudens(l), inchoatum (1-2) , innotatum (1-2), jucun- 
dum ( 1-2 ) , magister ( 1-2 ) , minus ( 1-2 ) , murchisoni ( 1-2 ) , particeps 
(1-2), pectinatum(l), poculum (1-2), pseudocalamiteum (1-2) , 
simois(l), socium (1-2), steiningeri (1-2), tenuicinctum(l) , tran- 
siens (1-2), trecentesimum(l) ; Phragmoceras callistoma (1-2), 
imbricatum (1) , panderi(l), perversum(l) ; Trochoceras amicum 
(1), nodosum (1-2) , placidum (1-2) , pulchrum (1-2?) . 

Gastropoda: Cirropsis bohemica Per. (1-2), G. disjunct a 
Per. (1-2), C. pvcEstans Per. (1-2), Conotoma eximia(l), Con- 
vadella inopinata(l), Cyvtolites evemita(l) y C. eximius(l), 
C. tuboides(l), Dyevia car ens (1), Epiptychia dusli Per. (1), 
Euvyzone tuboides Per.(l), Holopea ivvegidavis(l), Naticopsis 

11 In the latter period of his paleontological activity, Barrande used to 
refer the majority of the Silurian fossils imbedded in limestone to E 2 , although 
his collectors, workers in quarries, who brought the fossils to Prague, were 
getting them chiefly from E x limestones, which on the old weathered 
exposures yielded the fine, entire specimens with well preserved sculpture, 
such as are now' seen in the Bohemian Museum. Many such rich localities 
have disappeared or become nearly inaccessible; others yield from the fresh 
exposures chiefly fragments, hardly determinable, so that it is now a difficult 
task to ascertain the true horizon of all Barrande's species. 

18 The author is Barrande, unless otherwise stated. To save further space, 
E x and E 2 are hereafter written simply 1 and 2. 

Correlation of the Silurian of Bohemia. 65 

insculpta(l), Orthonychia ampla (1-2), 0. cuneus(l), 0. elegans 
(1-2), 0. nobilis(l-2), 0. togata(l), Platyceras cesopus (1-2) , P. 
complanatum (1-2), P. concors (1-2) , P. conviva (1-2), P. exsur- 
gens (1-2), P. for mo sum (1-2), P. forte (1-2), P. incho cms (1-2) , P. 
longipes (1) , P. cedematosum (1-2) , P. sanum(l-2), P. subcarina- 
tum(l), P. vexatum (1-2), Polytropis dives (1), P. pot ens (1), P. 
pulchra(l), P. robnsta(l), P. ventricosa (1) , Spirina patula (1-2) , 
S. tubicin a (1-2). 

Acephala: Astarte minuscida(l), Avicidopecten cybele(l), 
Avicula glabra(l), A. manulia(l), Cardiola expectans (1-2 ?), 

C. fortis(l-2), C. gibbosa (1-2), C. intermittens (1-2), C. inter- 
rupt a (1-2), C. migrans (1-2) , C. p etasina (1-2) , C. pidchella 
(1), C. radiat a (1-2), Dualina bella(l-2), D. branikensis (1) , 

D. extracta (1-2), D. fidelis (1-2) , D. humilis (1-2) , D. 
in con gruens (1-2), D. nympharum (1-2) , D. secunda (1-2) , 
Goniophora phrygia (1-2), Grammysia prcecox (1-2) , Maminka 
comata(l-2), M. tenax (1-2), Mila consanguis (1-2) , M. insolita 
(1-2), Modiolopsis inv olut a (1-2) , M. pupa(l), M. rebellis 
(1-2), Panenka capital a (1-2), P. gyrans (1-2), P. insocialis (1-2) , 
P. vendita(l), Paracardium complicatum (1-2) , P. delicatum 
(1-2), P. filiferum (1-2), P. incipiens (1-2) , P. mundum(l-2), 
P. rarissimum (1) , Posidonomya eugyra(l-2), Prcecardium 
adolescens (1-2) , P. b ohemicnm (1-2) , P. complacens (1) , P. 
fidens (1-2), P. ministrans(l), P. primidum (1-2) , P. quadrans 
(1-2), #Zai>a aberrans(l), S. bohemica (1-2). 

Brachiopoda: Atrypa insolit a (1-2) , A. sappho (1-2), A. 
s quamma (1-2) , A. thetis(l-2), Crania(l) bohemica(l), Discina 
nana(l), D. propinqua (1-2), Z>. imgim(l), D. truncat a (1-2), 
D. v exat a (1-2 1) , Eichwaldia bohemica (1), Lingula albicans 
(1-2), L. comes(l-2), L. dilatata(l), L. nigricans (1-2) , Orthis 
cognata(l), 0. honor at a (1) , 0. venustula(l), Pentamerus per- 
ditus(l), Rhynchonella niobe(l-2), Strophomena br act eola (1-2). 

Trelobita: Acidaspis dufrenoi (1-2%), A. mira(l), A. pre- 
vosti(l), A. rara(l), A. roemeri(l), Ampyx rouaidti(l), Are- 
thusina konincki(l), Bronteus p art schi (1-2) , B. planus Cda. (1), 
Cheirurus insignis Beyr. (1-2), Cromus beaumonti (1-2), Cyphas- 
pis burmeisteri (1-2), C. depressa(l-2%), Dalmanites orba(l), 
Deiphon forbesi (1-2?), Harpes naumanni (1-2%) , Illcenus bou- 
chardi (1-2), Lichas scabra Beyr. (1-2), L. palmata (1-2), Phacops 
glockeri (1-2) , Proetus decor us (1-2%), P. dentatidus Nov. (1), 
P. nasutus Xov.(l), Staurocephalus murchisoni (1-2) . 

Crustacea diversa: Aptychopsis prima (1-2), Ceratiocaris 
beraunensis (1-2), C. bohemica (1-2), C. decipiens (1) , C. grata 
(1), C. incequalis (1-2) , C. leptoglyphaN. (1), Cryptocaris pnl- 
chra(l), Discinocaris dusliana N. (1), Eurypterus acrocephalus 
S.(l), Holocaris univalvis N. (1), Plumulites minimus (1), P. 
squamatida(l) , Pterygotus beraunensis S.(l-2), P. barrandei N. 
(1), P. cyrtochela(l), P. hellichi N. (1). 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 19.— July, 1922. 

66 Perner & Kodym — Zonal Division and 

Diversa: Scyphocrinus decor atus W. & J.(l)', 8. excavatus 
(Schl.)(l), 8. subornatus(l), Bohemicocrinus pulv evens (1-2), 
Aulopora disjecta Poc.(l), Desmograptus plexus Poc. (1-2?), D. 
giganteus Jahn(l), Dictyonema bohemicum (1-2), D. confertum 
Poc. (1-2?), D. grande (1-2?), D. graptolithorum Poc. (1), Inocau- 
lis attrita Poc. (1). 

Bande F 

The highest member of the Silurian series in Bohemia 
is the dark limestones which have been referred by Bar- 
rande to his Etage F and designated as F^ Their litho- 
logic and faunistic characters place them much nearer to 
the E 2 beds than to the F 2 limestones, which are undoubt- 
edly of Lower Devonian age. 19 In addition, a faunistic 
hiatus between F x and F 2 , signifying the boundary line 
between the Silurian and Devonian in Bohemia, together 
with the decidedly Silurian aspect of the Fj fauna, points 
out the closer connection of Fj with E 2 . 

The boundary line between E 2 and F x is marked by the 
occurrence of hornstones in the limestone, and by a 
change of fauna. As in E x and E 2 , there are two facies 
developed in F x . The reef facies, which in F 1 has a far 
greater extension, is formed of limestones with very sub- 
ordinate shales. The limestones in the lower part of F x 
are dark, nearly black, fine-grained or compact, thickly 
bedded, and contain dark hornstones of irregular shape ; 
shaly reddish intercalations occur only in the lowest beds. 
In the upper part the limestones are lighter, gray or 
whitish, thinly bedded, finely or coarsely grained, resem- 
bling the white limestones of F 2 , but contain yellowish or 
reddish hornstones. Opposite Tetin this facies is fossil- 
iferous, bearing almost the same coral fauna as in E 2 . 

The other, "normal" facies is more shaly, and is lim- 
ited chiefly to a seam between Prague and Kosor (valley 
of Radotin). Characteristic are dark fine-grained lime- 
stones, splitting into flat plates and alternating many 
times with dark or brown shales ; hornstones are rare or 
completely missing. 

As for the fauna, it is quite different from the contem- 
poraneous coral fauna, and differs also from the "nor- 

19 It would perhaps be more appropriate to call the Fi beds E 3 , so that the 
Silurian of Bohemia would correspond to Barrande 's Etage E, and the letters 
F, G and H could be reserved for the Devonian. Some Bohemian geologists 
have already accepted this proposition. 

Correlation of the Silurian of Bohemia. 67 

mal" E 2 fauna ; a considerable part is, however, common 
to E 2 and F x . It is characterized by numerous species of 
Hercynella, which are regarded by some paleontologists 
as pelagic Pulmonata. 20 Moreover, cephalopods (40 
species, badly preserved, enumerated by Barrande and 
Novak), large lamellibranchs (Panenka, Prcducina, Modi- 
olopsis, Dalila, Cypricardinia, Hemicardium, Avicula), 
gastropods (St ropho stylus, Potellomphalus, Stylonema), 
Conularia, and Tentacidites, are other prominent groups ; 
a noteworthy feature is the thin shell of the molluscs men- 
tioned. Graptolites are represented by two last species, 
Monograptus hercynicus and M. kayseri Per. 

Concerning the more detailed division of these F 1 beds, 
I must content myself with giving a brief recapitulation 
of results published elsewhere, 21 and adding some further 
remarks. In the " normal " F x can be distinguished three 
horizons, which are not sharply delimited, as follows : 

1. The lowest consists of dark or black thick lime- 
stones, alternating with thick beds of shales ; it contains 
rare fish remains (M achcer acanthus , and some new genera 
resembling Ateleaspis, Aspidichthys, Cyathaspis, Dinich- 
thys, Macro pet alichthys, and Mylostoma). Other fossils 
are also rare. 

2. Fine-grained limestones, thinly bedded, alternating 
with shales. The greater part of the F x fossils occur 
here. Characteristic ones are: Hercynella nobilis, PL. 
radians, H. bohemica, H. paraturgescens, Rotellomphalus 
tardus, Strophostylus gregarius var. proeva, Stylonema 
solvens, Dalila obtusa, D. reseda, Lunulicardium analo- 
gum,L. evolvens, Panenka amcena, P. grata, Ceratiocaris, 
Aristozoe, Pygocaris, and Gigantostraca. 

3. Shaly limestones and shales (black, brown or gray) , 
with Monograptus hercynicus and M. kayseri. Spirifer 
inchoans is very abundant, and the following species are 
common: S. nerei, Atrypa canalicidata, Pentamerus 
janus, P. linguifer, Cyrtia trapezoidalis, Avicula migrans, 
A. pusilla, and Conocardium aptychoides. In the upper 
part of this horizon also appear trilobites. Common 
among them and very characteristic (confined to FJ is 

20 These are wanting in the reef limestone of F 2 (Koneprusy) and reappear 
in the shaly limestones of G^ 

21 Centralblatt f. Min. u. Geol., 19-20, 318-322, 1918. 

68 Perner & Kodym — Zonal Division and 

Bronteus umoellifer; others are rare and mostly frag- 
mentary. Their geological distribntion in other Silnrian 
and Devonian beds of Bohemia is very interesting and 
important, as may be seen from the following table : 

E 2 F x F 2 G-l G 2 

Acidaspis vesiculosa Beyr x x 

Bronteus umoellifer Beyr x 

Crotalocephalus gioous (Beyr.) x x x 

C rot aloe ephalus sternbergi (Boeck) x x x x 

Cyphaspis hydrocephala A. Roem x x x 

Harpes mieroporus Novak x 

Harpes venulosus Corda x x 

Phacops miser Barr x 

Proetus heteroclytus Barr x 

Proetus lepidus Barr x 

Proetus micropygus Corda x x 

We see that besides five species confined to ¥ l9 there are 
five which pass into the Lower Devonian (F 2 ) ; they are 
very rare in F l7 bnt common in the reef limestone of F 2 
(Koneprusy). It may be that the uppermost trilobite- 
bearing beds of this horizon belong already to the Lower 
Devonian. This possibility is also indicated by other cir- 
cumstances, such as the absence of the reef limestones of 
F 2 above the true Fj beds, which are in turn succeeded 
directly in many places by the upper division of F 2 beds. 
It would thus seem that these uppermost trilobite-bearing 
Fi beds are merely "normal' ' shaly deposits, contem- 
poraneous with the Lower Devonian white reef limestones 
of F 2 , a theory which might also explain the other great 
differences between upper F x and F 2 . 2 


22 Without dwelling longer on the Hercynian question in Bohemia, I will 
restrict myself here to the following remarks: (1) 0. Novak (Sitz. d. kgl. 
bohm. Ges. d. Wissensch., 1886) endeavored to prove that all of F t is equiva- 
lent to F 2 and therefore of Devonian age, a view which can no longer be sus- 
tained; (2) the Fi fauna exhibits a Silurian aspect (many species common 
to E 2 , frequent graptolites, absence of goniatites, see the f aunal list given by 
Novak and additions to it by zelizko (Verh. geol. Beichsanst., Wien., 9-10, 
1898) ) ; (3) the different opinions on F x and F 2 contained in the papers of 
Kayser, Holzapf el, Katzer, Freeh, Seeman, etc., are to be corrected in accord- 
ance with the researches of Kodym (Bohem. Acad. Sci., 1918), which prove 
that F 2 a (white reef limestone of Koneprusy) is a distinct lower horizon, 
succeeded by F 2J S (red marbles of Slivenec), the latter being identical with. 
Kayser 's and Holzapf el 's " Menaner-Facies ' ' of F, (crinoidal limestones). 

The surprising results of Barrois and Pruvost on the passage beds and 
boundary between Silurian and Devonian in the Calais coal basin, accepted 2 by 
Stamp for Shropshire and southern Wales, beeame known to me too late to 
be taken into consideration in discussing and correlating the F x beds in the 
present paper. 

Correlation of the Silurian of Bohemia. 69 

Correlation of Bohemian Strata. 

A detailed and definite parallelization of Silurian strata 
in Bohemia with those in foreign countries can not yet be 
made, therefore only a scheme of approximate correlation 
is inserted here, based chiefly on the geological distribu- 
tion of gr apt otitic and other faunas and on the zones con- 
nected therewith. Further investigations in Barrande's 
area will surely reveal many important details and solve 
questions either merely mentioned or omitted in the pres- 
ent connnunication ; moreover, future progress in zonal 
division and correlation in other countries will also 
modify the scheme here given. Because of limited space, 
two or three smaller districts of the chief countries are 
combined in single columns. The greatest changes will 
probably take place when future investigations are made 
with due regard to the principles, facts, and views dis- 
cussed in Ulrica's "Revision of the Paleozoic systems." 23 

To the scheme above, in which only the "normal" 
shaly facies and not the aberrant reef deposits are con- 
cerned, a few remarks can now be added. There is in the 
main divisions and succession of faunas in Bohemia a 
general harmony with those of other countries; in par- 
ticular, Eja, E^, and E^/ correspond well enough with 
the Rastrites-, Cyrtograptus-, and col onus -shales in 
Sweden. But in the subordinate zones, there are consider- 
able differences in their vertical limits, a feature not sur- 
prising when we consider the great geographic separation 
of the Bohemian basin and the varied development of its 
deposits (see for instance, the relative frequency of grap- 
tolites in limestones in association with various molluscs 
and brachiopods). The higher the zones, the greater the 
differences; this seems to be partly due to the fact, 
observed in graptolitic faunas in Great Britain by Miss 
E. M. R. "Wood, that graptolites of the highest Silurian 
zones show an extended vertical range and limited geo- 
graphic distribution. There seem to be considerable dif- 
ferences in the association of some graptolites in the 
higher Bohemian zones, as compared with those of "Wen- 

23 Bull. Geol. Soc. America, 22, 1911. I confess freely that Ulrich's work 
made so strong an impression upon me that I hesitated to publish the present 
communication and was about to review all my former studies from the 
standpoints of that work. This would, however, cause another considerable 
delay, so that I preferred to print the present article (cf. footnote 6), the 
purpose of which is to offer at once a general orientation to foreign students. 

70 Perner £ Kodym— Zonal Division and 

Correlation Table op 


England and Scotland 

(Lap worth) 

(Elles, Wood, Lake) 

f i 





S 2 






Upper Ludlow 


M. testis 












1 CO 
CO fl 
fl CO 
05 ft 
fcl TJ 
+> CO 
• ti 



e lY 


M. vulgaris 








C. symmetricus 
M. riccartonensis 




C. murehisoni 




C. murehisoni 



e i« 



Op. Llandovery 



M. gregarius 


D. vesiculosus 

Low. Llandovery 











D. anceps 


iny,Trinucleus sh.) 

Correlation of the Silurian of Bohemia. 


European Silurian 

Sweden, Skane, 


2 03 

O fad 


(Bergeron, Barrois, 


TJ +» 
03 «■* 
r-4 B 
J£ JZ 
+> O 
63 -w- 



Ascoceras beds 

Megalomus and 











Sh.and qu. of 











(Slava bohemica 







7(Bolbozoe bohemica 





( soni,Hy. simplex 


M. testis 

( don, C.rigidus 























M. spiralis 

M. spiralis 



1 ' II 

a t> 

03 >> 

r-l ti CO 

• > 03 

a o +» 

□ -8 B 



o co 

03 W 
r-4 03 

J -r» 
• 63 

u3 a 



G 3 


C". comet a 





_Ce_ptuf oliM 







W~. cy ph as 







ichiopoi shales 
3d D.anceps beds 





Dpper Ordovician 
(Orthis actonlae, 

72 Perner $ Kodym — Silurian of Bohemia. 

lock or Lower Ludlow age in Great Britain, or in cor- 
responding strata in different Scandinavian districts ; but 
as the limits of higher zones in all countries are not as 
sharp as those of the lower ones, and hence the range of 
many of the youngest graptolites is not everywhere coin- 
cident, it would exceed the scope of the present communi- 
cation to dwell longer on the particulars. Interesting 
data along this line may be found in papers by J. Marr, 24 
Lake, 25 E. M. R. Wood, 26 and G. Elles, 27 as regards the 
British Silurian, and in the papers of Tullberg (1. c), 
Munthe, Warburg, Hedstrom, Wiman, Moberg, 28 and 
Kiaer 29 for Scandinavian districts. 

In spite of this, however, the graptolites are, generally 
speaking, more reliable zonal guides than other groups in 
these beds. Thus, for instance, the world-wide Cardiola 
interrupt a appears already in E^ and survives until the 
highest E 2 beds, while in other countries it is recorded up 
to much higher beds than those of EjjS, and has a more 
limited range. Further discussion on the correlation of 
Bohemian beds will be deferred until more facts shall 
have been gathered by new investigations. 

Prague, Czekoslovakia, July, 1921. 

24 Geol. Mag, 534, 1892. 

25 Quart. Jour. Geol. Soc, London, 51, 22, 1895. 
28 Ibid., 56, 415, 1900. 

27 Ibid., 56, 370, 1900. 

28 For the five preceding authors see Guides for Excursions, Xlth Internat. 
Geol. Cong., Stockholm, 19-22, 40, 1910. 

29 Videnskabs-selsk. skrifter, 2, Kristiania, 1908. 

Chemistry and Physics. 73 

I Chemistry and Physics. 

1. A New Process for the Industrial Production of Barium 
Hydroxide for the Treatment of Molasses in Sugar Refining. — 
Since the discovery of the sparingly soluble barium sucrate by 
Peligot in 1838 many attempts have been made to extract the 
cane sugar from molasses by the precipitation of this compound, 
but up to the present time the cost of converting the barium car- 
bonate produced in the process into the hydroxide has been too 
great to render the operation remunerative, although it is stated 
that in the French beet-sugar industry about 15% of the sugar 
fails to crystallize directly. Deguide and Bode have now found 
a promising method for carrying out this process. They ignite 
the barium carbonate with an addition of silica in the proper pro- 
portion to yield 3BaO.Si0 2 . This mixture does not fuse at the 
temperature of about 1300° C. necessary for the transformation, 
and when the product is treated with water a very large propor- 
tion of the barium goes into solution as barium hydroxide. The 
process has been tried on a rather large scale, but has not yet been 
put into commercial operation. The authors state that the 
mother-liquors from the filter-presses containing barium sucrate 
allow the recovery of nitrogen and potassium taken from the soil, 
representing 1% and 6%, respectively, of the molasses employed. 
Comptes Rendus, 174, 1177. h. l. w. 

2. An Advanced Course of Instruction in Chemical Princi- 
ples; by Arthur A. Noyes and Miles S. Sherrill. 8vo, pp. 
310. New York, 1922 (The Macmillan Company).— This is a 
very notable and unusual textbook of advanced physical chemis- 
try. Instead of presenting the subject in the usual descriptive 
manner, the course is so planned as to give an intensive training 
in the application of the principles of the science to concrete 
problems. The text is interspersed with numerous problems, 
which generally require clear, logical thinking and a thorough 
understanding of the principles for their solution. These prob- 
lems are the important feature in the course of instruction. 

The subjects treated in the course have been selected on 
account of their fundamental and practical importance to chem- 
ists. For certain reasons, chiefly to avoid making the course of 
study too long, some topics such as radiation, atomic structure, 
colloidal solutions, etc., have not been included. The course of 
study laid out is evidently not an easy one for the student, but 
it should be of much value in developing his reasoning-power as 
well as in equipping him for a career in educational, research, or 
industrial chemistry in which the principles of physical chem- 
istry may be applied. It is the opinion of the authors that 120- 

74 Scientific Intelligence. 

150 exercises are required for covering satisfactorily the whole 
course as presented, but a provision is made for employing the 
book for a shorter course by the omission of certain designated 
articles and problems that are less important or more difficult 
than the others. h. l. w. 

3. The Determination of Sulphur in Iron and Steel; by H. B. 
Pulsifer. 8vo, pp. 160. Easton, Pa., 1922 (The Chemical 
Publishing Company). — The determination discussed in this 
book is an exceedingly important one, since sulphur has a delete- 
rious effect upon the metals even in small amounts, and small 
variations in the amounts present may have important effects 
upon quality. As the total amount of sulphur present, except 
in certain crude pig-irons, is usually less than 0.100%, and often 
below 0.010%, the analytical problem is a delicate and difficult 
one. A vast amount of research has been devoted to methods 
of making this determination, and an excellent feature of the 
book under consideration is an extensive and evidently very com- 
plete bibliography of the subject, with many interesting com- 
ments and citations of results. The bibliography covers prac- 
tically the whole period of modern chemistry, as it begins with 
the year 1797. It takes up about two-thirds of the book, and it 
is not only interesting historically, but it should be of great value 
to future workers on this problem. 

The author has made and records here a very large number of 
sulphur determinations by several methods, and he recommends 
the use of an "evolution" method, where the sulphur is first 
converted into hydrogen sulphide by the action of concentrated 
hydrochloric acid upon the metal in an apparatus of his own 
modification, then the hydrogen sulphide is absorbed in an 
ammoniacal solution of cadmium chloride, and after acidifying 
the latter the sulphur is determined by titration with iodine solu- 
tion. It does not appear to the reviewer that the author has 
definitely shown that his preferred method is the most reliable 
one, and it must be admitted that his results by different methods 
show unexpectedly wide variations. h. l, w. 

4. Organic Chemistry; by Victor von Richter. Vol. II. 
Chemistry of the Carooeyclic Compounds. Translated from the 
11th German edition by E. E. Fournier D'Albe. 8vo, pp. 760. 
Philadelphia, 1922 (P. Blakiston's Son & Co.).— This transla- 
tion appears ten years later than the corresponding German 
edition, but it is to be heartily welcomed as a work of much 
importance to English-reading students of organic chemistry. 
For many years the various editions of Richter 's Organic Chem- 
istry have been very valuable sources of study and reference to 
advanced students in the subject, as they have given excellent 
presentations of the theories, as well as descriptions of very large 
numbers of compounds. Our older chemists will remember that 

Chemistry and Physics. 75 

the earlier English translations of the work were made in a very 
satisfactory way by Dr. Edgar F. Smith of the University of 

At present the work appears in two volumes, the first of which 
deals with the aliphatic compounds. The recent German edi- 
tions have been prepared under the direction of Dr. Richard 
Anschutz of Bonn. h. l. w. 

5. Friction and Lubrication. — In discussing lubrication two 
cases are to be distinguished : (1) When two surfaces are floated 
apart by a lubricant, static friction is absent and the resistance to 
motion varies directly as the viscosity of the lubricant. ( 2 ) Where 
two solid surfaces are near enough together so that the friction 
depends not only upon the lubricant but also upon the chemical 
nature of the surfaces, the resistance varies as some inverse func- 
tion of the viscosity of the lubricant. This second case is termed 
boundary lubrication. It has recently been studied by Hardy 
and Doubled ay using polished surfaces of steel, glass, and bis- 
muth lubricated with liquids of the paraffine series. To assure 
definiteness of contact area, one of the surfaces was made of 
spherical form, and the other a plane. The coefficients of friction 
were first determined for the "clean" state where all possibly 
removable impurities had been abstracted. This is to be judged 
from the fact that the friction has a high and constant value. 

The variables studied were (a) the weight of the slider, (b) the 
curvature, (c) the thickness of the layer of the lubricant, (d) the 
chemical nature of the lubricant and the solids respectively. The 
authors find that the friction is strictly proportional to the weight 
or that the coefficient of friction is (a) independent of the weight 
and (b) of the curvature. In studying (c) three methods of 
lubrication were adopted: (1) the flooded state, where the 
slider stood in a pool of the fluid; (2) the primary film. When 
a drop of lubricant with sensible vapor pressure is placed on a 
clean plate, although the drop to all appearance remains where 
placed without change, an invisible film nevertheless spreads so 
as to cover the whole plate as is evidenced by the fall in friction ; 
(3) lubrication by deposit from the saturated vapor alone. 

The results showed that the friction was independent of the 
quantity of lubricant present provided there was enough to cover 
the surfaces with the invisible primary film. Where the amount 
was less than this critical value the fall of friction was propor- 
tional to the concentration of the molecules in the gaseous phase, 
indicating that each molecule exerts the same influence as every 
other. The influence of chemical constitution (d) was found to 
be unexpectedly simple. If /jl denote the coefficient of friction 
and M the molecular weight, their relation is given by 

^— b—a M 

where a and b are parameters. The parameter a is independent 

76 Scientific Intelligence. 

of the nature of the solid faces and depends only on the chemical 
type of the lubricant, varying from one chemical series to 
another. The parameter b however depends both upon the 
chemical series and the nature of the solid face. The interpreta- 
tion of these results is difficult. The authors do not accept the 
current view that friction is due to interlocking asperities of 
molar dimensions but rather depends upon molecular attraction 
across an interface. If asperities are to be considered they must 
be the atoms and the molecules of the substances. The effect of 
a tangential force is not merely to move the atoms and molecules 
in the tangential plane but also to rotate them. 

The mental picture which the authors offer is that the primary 
film of lubricant consists of a single layer of molecules which 
have been oriented by the attractive fields of the solids so that 
their long axes are perpendicular to the solid faces. When two 
such surfaces are brought together the friction represents the 
maximum tangential stress which can be supported at some 
median plane of slip which is an imaginary surface parallel to 
the surface of the solid. This resistance is obviously conditioned 
by the attractive field of the solids which must vary rapidly along 
the normal, and also by the nature of the molecular chain of the 
lubricant used. — Proc. Roy. Soc. 100, 550, 1922. f. e. b. 

6. Power Alcohol; by G. W. Monier- Williams. Pp. xii, 323. 
London, 1922 (Henry Frowde and Hodder & Stoughton). — In 
view of the possible insufficiency of the world's petroleum 
reserves the investigation of alternative motor fuels is an 
important subject which is very fully examined in this book. 
The first chapter passes in review the questions of the supply, the 
production and the economy of the various motor fuels, and 
reaches the conclusion that there is strong probability that before 
many years the supply of gasolene will be permanently unequal 
to the demand, and that power alcohol has an undoubted future 
before it as a supplementary if not as a competitive fuel. Chap- 
ter 2 discusses the way in which various organic constituents of 
plants are elaborated, and the chemical processes by which sugars 
are converted into alcohol by the yeasts. Chapter 3 treats of 
the raw materials from which fermentation alcohol is produced, 
namely, plants supplying starch, plants containing ready formed 
sugars, and cellulose, and also of the commercial processes of 
mashing, saccharification, fermentation and distillation. Chap 
ter 4 surveys the more important starch or sugar-bearing raw 
materials from the point of view of yield, availability and related 
economy of alcohol production. Chapter 5 is devoted to the 
commercial treatment of cellulose materials for the production of 
alcohol. Chapter 6 explains the ways of making synthetic alco- 
hol and its production on a commercial scale. 

Chapter 7 reviews the methods of denaturation and various 

Chemistry and Physics. 77 

matters in connection with excise supervision in Great Britain. 
Chapter 8 discusses the principles of the internal combustion 
engine and its efficiency. In Chapter 9 the chemical and physi- 
cal properties of alcohol are presented in full detail both by tables 
and by diagrams. Chapter 10 summarizes with considerable 
detail the results of alcohol-engine tests with regard to per- 
formance and efficiency. These compare very favorably with 
gasolene-engine tests except in the matter of fuel consumption. 
Apparently the price of alcohol will have to fall considerably 
below that of gasolene before it can be regarded as a commercial 

Another and more promising line of attack on the motor fuel 
problem is in the direction of mixtures of alcohol with other 
easily volatile substances such as ether or benzol, or even with 
gases such as acetylene. 

The final chapter (11) discusses various proposals of this 
nature but the subject is too extensive to receive adequate treat- 
ment in a book of this character. Appended to each chapter will 
be found a useful and numerous list of references to the particu- 
lar topics which have been under discussion. f. e. b. 

7. The Journal of Scientific Instruments.— The Institute of 
Physics (England) in connection with the National Physical 
Laboratory proposes to publish a journal devoted to the theory, 
construction and use of instruments as an aid to research in all 
branches of science and industry. The preliminary number 
appeared in May. To assure its continued existence a subscrip- 
tion list of about 3,000 is desired. f. e. b. 

8. La Theorie Einsteinienne de la Gravitation; by Gustave 
Mie, translated from the German by J. Rossignol. Pp. xi, 118. 
Paris, 1922 (J. Hermann). — This little volume contains a clear 
and readable exposition of the relativity theory and Einstein's 
theory of gravitation couched in simple language and free from 
mathematical symbols. It should appeal to the reader who has 
a knowledge of the principles and laws underlying the subject of 
physics, but who is not conversant with the somewhat compli- 
cated differential geometry involved in a detailed presentation of 
the general relativity. The author 's emphasis on the philosophi- 
cal aspects of the subject should make the book particularly 
interesting to those philosophers who are looking for a clear and 
simple account of the revolutionary changes in our concepts of 
space and time which the relativity principle has occasioned. 

A short mathematical appendix (11 pages) added by the trans- 
lator presents in condensed form the author's derivation of 
Einstein's law of gravitation. l. p. 

78 Scientific Intelligence. 

II. Geology and Mineralogy. 

1. Gravity Anomalies and their Geological Interpretation*. — 
This article is in part a condensation of a more detailed paper by 
the same author entitled "Die mediterranen Kettengebirge in 
ihrer Beziehung zum Gleichgewichtszustande der Erdrinde." 
(Abh. Sachs. Akad. d. Wiss., Bd. 38, 2. Leipzig, 1020.) The 
article opens with some general explanations of the formulas for 
theoretical gravity and of the methods for reducing gravity 
observations for topography. This opening section contains two 
or three statements not strictly correct, but these concern mat- 
ters not generally understood except by specialists in gravity 
work or in the theory of the figure of the Earth, and do not 
essentially affect the conclusions reached. 

The Bouguer anomaly is used throughout the article as a 
measure of the compensation. A single Bouguer anomaly meas- 
ures, of course, merely the net effect of the compensation and of 
the distant topography and affords no clue to the mass of the 
compensation nor to its situation. Thus no allowance is made 
by this method for the effect of nearness to the continental shelf, 
but for nearly all of Europe this effect is practically negligible. 
The advantages of using Bouguer anomalies are the ease with 
which they are computed and the fact that they imply no par- 
ticular depth or distribution of compensation, and that, when 
gravity stations are sufficiently dense, they lend themselves, 
according to Prof. Kossmat, to geological interpretation better 
than the anomalies computed by the isostatic method. 

The article summarizes the results of Prof. Kossmat 's examina- 
tion of known gravity anomalies in Europe, Asia, Africa and 
North America. For the last-named the war has evidently pre- 
vented him from receiving the more recent publications on the 
subject. He maintains that mountain chains are not caused by 
a swelling of the crust in their vicinity but by the folding under 
tangential pressure of certain weaker portions of the crust. 
Such a chain is not compensated by itself but only in connection 
with the neighboring piedmont regions. The additional matter 
imposed upon a given area by the folding may be partly com- 
pensated by the sinking of that area under the additional load. 
This sinking the author conceives as partly elastic rather than 
entirely flotational, the substratum being bent under the load. 
The result is that the compensation is regional rather than local. 
In regard to the evidence on this subject offered by the 124 sta- 
tions in the United States, the number treated in U. S. Coast and 
Geodetic Survey Special Publication No. 12, which evidently 

* Die Beziehungen zwischen Schweranomalien und Bau der Erdrinde, by 
Franz Kossmat. Geologische Eundschau, vol. 12, pp. 165-189. 1921. 

Geology and Mineralogy. 79 

was the latest publication available to him, Prof. Kossmat is non- 
committal. In accepting the existence of large tangential forces 
he does not accept Wegener's ideas as to the crustal movements 
produced by them. At the close of the article there is a useful 

Geologists have been rather slow to utilize the evidence avail- 
able from gravity observations in regard to the distribution of 
density in the Earth's crust, evidence which began accumulating 
when Bouguer from 1736 to 1740 made pendulum observations 
in what is now Ecuador. One reason for the neglect of this 
accumulating geodetic evidence has doubtless been the difficulty 
of finding an unequivocal interpretation, but as gravity stations 
and stations where the deflection of the plumb line is known 
become more numerous and more densely distributed over a 
given area, the practical range of our interpretations of this 
evidence in terms of density, if not the range of purely specula- 
tive mathematical possibilities, becomes more and more restricted 
and the evidence of more real service to the geologist. Articles 
like the present are a welcome sign that geologists are beginning 
to realize that geodesy has evidence of value to them and that 
compensation for every topographic feature exists in some degree 
and must be explained by the adopted hypotheses of dynamic 
geology. Both geologists and geodesists will find the article 
profitable reading even though they may not accept all of the 
author's conclusions. w. d. lambert. 

2. Publications of the United States Geological Survey, 
George Otis Smith, Director. — Recent publications of the U. S. 
Geological Survev are noted below: (See earlier, 3, 97-98, Jan., 

Axnual Report: — Forty-second Report of the Director for 
the year ending June 30, 1921. Pp. 108, 1 plate. — This publica- 
tion gives in detail the work carried through by the Survey dur- 
ing the year noted. The Director dwells in particular upon the 
economy attained through a higher degree of efficiency. The end 
sought for is one greatly to be desired in all departments of the 
Government, but, unfortunately not often attained. In the case 
of the Survey the efficiency striven for is much limited by con- 
ditions which have already existed too long. These include 
inadequate office quarters, restriction in the selection of person- 
nel, small salaries and reduced appropriations for printing. The 
total funds available for 1920-21 were $105,575. The authorized 
expenditures were distributed as follows : For economic geology 
of metalliferous deposits, $50,000 ; of non-metalliferous deposits, 
$23,575; of oil, gas, coal, $110,000. Scientific researches not 
directly economic were allotted $117,000 ; while administration, 
salaries, etc., received $105,000. Of the total sum available for 
geologic work, $122,000 was used for field expenses (including 

80 Scientific Intelligence. 

search, for potash) ; of this 75 per cent was expended west of the 
one-hundreth meridian. 

Other publications are : 

Professional Papers, No. 123. — A Superpower system for the 
region between Boston and Washington ; by W. S. Murray and 
others. Pp. 261, 11 pis., 61 text figures. This is the report of a 
special investigation of the possible economy of fuel, labor, and 
materials resulting from the use in the industrial region named 
of a comprehensive system for the generation and distribution of 
electricity to transportation lines and industries. The investiga- 
tion was made by a staff of engineers under the administrative 
supervision of the Geological Survey The annual net saving, 
after fixed charges are deducted, if the energy required in this 
region in 1930 were supplied by a co-ordinated power system, as 
described, is estimated at $429,000,000 ; the annual return on the 
investment would be 33 per cent. Of the 36,000 miles of railroad 
in this region it is estimated that 19,000 miles can be profitably 
electrified, so as to yield by 1930 an annual saving of $81,000,000 
as compared with the cost of operation by steam. The coal saved 
annually under the superpower system would amount to 50,000,- 
000 tons. The system contemplates interconnection of existing 
plants and systems and construction of new steam-electric and 
hydroelectric plants at the most favorable locations. This 
"superpower zone" embraces one-fourth of the population of 
the United States, with a corresponding importance in railroads, 
industrial plants, etc. 

No. 129. Parts C to I. 

Topographic Atlas : — Forty-three sheets. 

Geologic Folio, No. 213. New Athens-Okawville, Illinois ; by 
E. W. Shaw. 12 pages of folio text, 4 maps, 6 text figures. 

Bulletins : — Several parts, separately issued, of Nos. 722, 725, 
726, 730, 735. Part B of No. 735 by L. F. Noble describes the 
colemanite of Clark County, Nevada. This occurs in fairly 
extensive deposits of economic importance. 

Water Supply Paper: — No. 500, C. Some characteristics of 
Run-off in the Rocky Mountain Region ; by Robert Follansbee. 
17 pages, 10 text figures, and in 

Mineral Resources. — Numerous advance chapters. 

3. Die Eruptivgesteine des Kristianiagebietes IV. Das Fen- 
gebiet in Telemarh, Norwegen; by W. C. Brogger. Videnskap. 
Skrifter. I. Mat.-naturw. Klasse 1920, no. 9. 408 pages, with 2 
geologic maps, 30 plates, and 46 text figures. Christiania, 1921. 
— It is nearly a quarter of a century since the third volume of the 
epoch-making series "Die Eruptivgesteine des Kristianiage- 
bietes" appeared. During this interval detailed mapping of 
the region has been in progress, and a vast amount of material 

Geology and Mineralogy. 81 

lias; been accumulated — about 15,000 hand specimens, 6,000 thin 
sections, and 300 chemical analyses. The results of this work 
were to have been embodied in volume IV, but the discovery of a 
remarkable series of alkali rocks in the Fen district by Professor 
Brogger's colleague, Professor V. M. Goldschmidt, in 1918 made 
it imperative to study this district at once, because it differs so 
strongly from the Christiania district. It is the results of this 
study, this interim investigation as it were, that are presented in 
this superb monograph. It need hardly be said that it represents 
a very notable achievement, but coming as it does at the time of 
Professor Brogger's seventieth birthday, the rernarkableness of 
the achievement is enhanced many fold. 

The outstanding feature of the Fen area is the occurrence of 
large intrusive masses of limestone and dolomite in association 
with ijolitic rocks that range from highly leucocratic to highly 

The Fen area, lying 115 km. southwest of Christiania, is 4.2 
square kilometers in extent, and is surrounded on all sides by 
Pre-Cambrian granite. A little less than half of the area con- 
sists of intrusive limestone, termed sovite. The predominant 
rocks consist of a series of plutonic pyroxene-nephelite rocks. 
In composition they range from hypermelanocratic members — 
jacupirangite — through members in which pyroxene predom- 
inates (melanocratic rocks here named melteigite), through 
members in which nephelite predominates (the leucocratic rock 
ijolite), finally to the hyperleucocratic end member of the series 
— urtite. All these rocks carry more or less primary calcite. In 
connection with the systematic description of this series, the 
occurrence of all ijolitic rocks the world over is reviewed and 
critically discussed, and melteigite is shown to occur in various 
alkalic provinces. 

At Fen melteigite predominates. The melteigite magma 
exerted a truly remarkable and profound contact metasomatism 
on the Pre-Cambrian granite, the result of this action being to 
transform the bordering zone into a wide band of aegirite sj^enite, 
or fenite. Fenite has also been formed by the assimilation of the 
granite by the melteigite magma. Chemically and texturally 
the pyrogenetic fenite often cannot be distinguished from the 
fenite of metasomatic origin. 

Countless dikes of calcite (sovite) and a lesser number of 
dolomite (rauhaugite) cut the melteigite, ijolite, and fenite. 
Their occurrence in the fenite of metasomatic origin rules out 
the possibility that the dikes may in reality be roof -pendants. 
Professor Brogger is emphatic that the carbonate rocks are truly 
igneous and leaves us in no doubt as to what he means by a dike, 
namely that a dike consists of material that was injected as a 
homogeneous fluid mass during one pulse, filling the fissure and 

Am. Jour. Sci.— Fifth Series, Vol IV, Xo. 19.— July. 1922 

82 Scientific Intelligence. 

subsequently consolidating there. The dikes are described as 
streaky (schlierig) and as unusually rich in apatite, containing 
as much as 8 per cent, fluidally arranged. These nuxional struc- 
tures would appear to be the strongest evidence in support of the 
igneous origin of the carbonate dikes. Further evidence is 
afforded by the complete series of transitional rocks from pure 
silicate rocks (ijolites) to pure carbonate rocks. Many of these 
intermediate rocks have excellent "eutectic" structure, which is 
regarded as sure proof of their igneous origin, but in view of the 
fact that in recent years many of the so-called eutectic inter- 
growths have been shown to be due to replacement, this criterion 
has not the absolute value that the author attaches to it. 

From the petrographic resemblance of the central mass of 
limestone of the Fen area to the sovite dikes, it also is interpreted 
as of igneous origin. The possible length of the limestone area 
is 2 km. and its width is 1 km. The intrusive carbonatites and 
older rocks are cut by aliio'itic dikes, and all are intruded by 
diabase dikes, but these last injections are not believed to have 
any genetic relation with the Fen magma and its consanguineous 

Daly's hypothesis for the origin of alkaline rocks is appre- 
ciatively considered, but it is believed that at Fen the parent 
magma was essexitic rather than basaltic. Igneous activity 
began at Fen by the formation of a volcanic conduit by a gigantic 
explosion. The essexite magma dissolved at great depth large 
quantities of limestone, thereby producing the melteigite magma, 
which, by gravity differentiation eventually yielded urtite. 
After differentiation the carbonate magma swam on top of the 
heavier silicate magma and finally solidified in the throat of the 
volcano. Subsequently the volcano has been very deeply eroded, 
and to this great denudation is ascribed the marked difference 
between the igneous phenomena revealed at Fen and those in the 
Christiania region : at Fen the phenomena are those due to deep- 
seated igneous activity, whereas in the Christiania region they 
are those produced in the comparatively shallow depths of the 
earth's crust, adolph knopf. 

4. Mineral Resources of the Philippine Islands for 1919 and 
1920. Pp. 70, with 4 plates and 4 text figures. Manila, 1922.— 
The mineral production of the Philippines has increased from 
234,000 pesos in 1907 to 7,611,000 pesos in 1920. In money value 
gold is much the most important, namely 3,000,000 pesos (1916), 
2,620,000 (1919), 2,425,000 (1920.) Coal increased from nearly 
400,000 pesos (1918) to over 800,000 {1919) and 1,452,000 
(1920). Salt, sulphur, stone and sand, with other products, are 
also important. Iron ore exists though not yet developed and 
the existence of a high grade of petroleum is believed to be 
assured when active work of prospecting can go on. 

Geology and Mineralogy. S3 

5. A List of new Crystal Forms of Minerals; by Herbert P. 
Whitlock. Bull. Amer. Museum Nat. History, vol. 46, pp. 
89-278— Dr. Whitlock in 1910 published a list of new crystal 
forms in the School of Mines Quarterly (31, p. 320, 32, p. 51) ; 
this included forms described subsequent to Goldschmidt 's Index 
(1891). The same author has now in this pamphlet of about 190 
pages covered the entire period from 1891 to 1920, and his ardu- 
ous labors will be highly appreciated by all interested in this 
subject, one that no work on mineralogy, however comprehensive, 
can expect to cope with. The accepted letter, the Goldschmidt 
and Miller symbols, locality, and original reference are all given. 

6. Handbook and Descriptive Catalogue of Gems and 
Precious Stones in the U. S. National Museum; by George P. 
Merrill, assisted by Margaret \V. Moodey and Edgar T. 
\Vherry. Pp. 225, with 11 plates, 26 text figures. (Bulletin 
118, U. S. National Museum.) — Dr. Merrill's contribution will be 
welcomed by all interested in precious stones, particularly those 
of American origin. The collection, even if ' ' poorly balanced, ' ' 
is large and worthy of careful study. The "Isaac Lea Collec- 
tion" forms a very important part; this originated with the 
bequest of Mrs. Frances Lea Chamberlain of the collection of her 
father, Dr. Isaac Lea. This has been made more valuable by the 
further gift from Dr. L. T. Chamberlain of numerous specimens 
and (on his decease) a bequest for the increase of this collection. 

7. Virginia Geological Survey; Thomas L. \Yatson, Director. 
Bulletin XVII. The Geology and Coal Resources of Russell 
County; by Chester K. Wentworth, with a chapter on the For- 
ests of Kussell County by J. W. 'Byrne. Pp. X, 179 ; 28 plates 
(3 in pocket), 16 figures. Charlottesville, 1922. — Russell County 
is in the southwestern part of the State and the part here 
described is a belt, 3 to 5 miles wide, along the northwestern 
border. Its resources are estimated at 706,000,000 tons of recov- 
erable coal ; in 1918 the output was very nearly 2,000,000 tons, 
having a value of over $5,000,000. 

8. The Topographic and Geological Survey of Pennsylvania; 
George H. Ashley, State Geologist.— Bulletins 6, 9, 38-42 have 
been received (mimeographed). Xo. 40, by F. B. Beck, is on the 
white clay deposits at Saylorsburg, Monroe Co.; the others (see 
3, 305, 384) deal with various aspects of the coal situation. 

9. Geology of Drumheller Coal Field, Alberta ; by John A. 
Allan. Pp. 72 with index; 17 plates including a large colored 
geological map of the Drumheller district in pocket. — This dis- 
trict is the largest producer of coal in Alberta. Within the 75 
square miles mapped there are 28 mines and the annual output is 
over one million tons. The coal seams and associated sedimen- 
tary beds belong to the Edmonton formation, the uppermost 
member of the Upper Cretaceous. The area is 8o miles east- 

81 Scientific Intelligence. 

northeast from Calgary and 185 miles south of Edmonton. It 
is assumed that in the Red Deer and tributary valleys within the 
area mapped there is an aggregate of ten feet of workable coal 
over at least 6,000 acres ; this is figured as meaning a reserve of 
over one hundred million tons. 

10. Potash in a new area of Texas, deposit of polyhalite found 
at several levels in a well in Reagan County, Texas. — The dis- 
covery of potash in notable amounts in a new area in Texas is 
announced by the U. S. Geological Survey. This has been 
brought to light through the analysis by the Survey of drill cut- 
tings collected from the Santa Rita No. 1 well, drilled by the 
Texon Oil and Land Co., in the southwest corner of Reagan 
County. Most of the samples contained no potash worth noting, 
but the samples which were taken from bailings at depths of 1150 
to 1325 feet showed from 2.05 to 8.29 p. c, K 2 0. The richest of 
the samples indicates 10.78 per cent of K 2 in the soluble salts 
when one gram of the dry rock is dissolved in 100 cc. of water. 

III. Natural Htstorv. 

1. Arctic Alcyonaria and Actinaria; I, II, The Alcyonaria 
of the Canadian Arctic Expedition, 1913-1918, with a revision of 
some other Canadian genera and species; III. The Actinaria of 
the Canadian Arctic Expeditions, with notes on interesting 
species from Hudson Bay and other Canadian localities; by 
A. E. Verrill. Pp. 161, with 31 plates. Ottawa, 1922.— These 
papers from volume VII of the Report of the Canadian Arctic 
Expedition, 1913-191S, contain a full systematic revision, with 
specific descriptions and illustrations, of all the species of these 
groups known from Arctic North American waters, including 
some of the more characteristic forms from the Grand Banks. 
Several new genera and species, are included, and the previously 
perplexing confusion in synonymy is eliminated. 

Zoologists will long be indebted to Professor Verrill for bring- 
ing together in this monographic form the accumulated results 
of his studies for more than half a century on these groups. 

w. r. c. 

2. Genetics: An Introduction to the Study of Heredity: by 
Herbert Eugene AValter, Revised edition ; pp. xvi, 351. New 
York, 1922 (The Macmillan Company). — In the ten years that 
have passed since the first edition of this widely used text-book 
was issued discoveries of such profound significance have been 
made as to necessitate extensive revision and the rewriting of a 
large portion of the work. Several new chapters have been 
added and a number of cleverly elucidative diagrams incorpo- 
rated. The first edition was recognized as one of the most suc- 
cessful elementary treatises of the subject of genetics that have 

Natural History. S5 

been written since the discovery of the Mendelian principles. 
Its usefulness is attested not only by its wide adoption as a text- 
book in English-speaking countries but also by its translation into 
other languages. 

The present revision will restore the book to its former pres- 
tige. In it both the student and the general reader may find the 
essential facts of the rapidly advancing science of heredity pre- 
sented in a most orderly, intelligible, and discriminating manner. 

w. k, c. 

3. A Naturalist in the Great Lakes Region; by Elliot Row- 
land Dowxixg. Pp. xxv, 328, with 452 illustrations. Chicago, 
1922 (The University of Chicago Press). — The object of this 
book is to place in the hands of school children an intelligible 
and inspiring story of the natural world about them. The first 
chapter shows them the unfinished world at their* feet, with the 
elements ever at work changing the topography of hill and valley 
— filling up the ponds and eroding the mountains. They next 
read about the world in the making and the story told by the 
rocks and by the boulders in the fields as to the changes which 
have taken place on the surface of the earth in its past history. 
Then the living world is surveyed and the adjustment among the 
different groups of organisms is pointed out. This is followed 
by visits to the dunes, the ponds, the swamps, forest, prairie, 
ravine, brook, and river. In each of these pleasant and instruc- 
tive excursions the reader makes the personal acquaintance of the 
more interesting plant and animal inhabitants of the regions and 
learns the reasons why each has become settled in its particular 
environment. Although the modern name for this method of 
studying nature may be ecology, it is, nevertheless, good old- 
fashioned natural history. 

The book is most attractively bound in flexible covers and will 
go easily into a coat pocket, and it should be carried afield not 
only by the pupil in nature study courses, but also by grown-ups 
whenever they are fortunate enough to have an hour to spend 
with nature. w. r. c. 

4. La Constitution des plantes vaseulaires revelee par leur 
Ontogenie; by Gustave Chauveaud. Pp. xiii, 155, with 54 text- 
figures. Paris, 1921 (Payot & Cie.). — In this interesting 
pamphlet the author explains and defends his theory of the 
' ' phyllorhiza," ' according to which the f unclameiital unit of the 
vascular plants is neither the stem nor the leaf but a composite 
organ composed of a leaf-like part united to a root-like part. In 
certain aquatic pteridophytes these units are clearly shown by 
young embryos; the first phyllorhiza gives rise to a second by 
means of a vegetative point arising laterally, the second gives 
rise to a third, and so on. At first the successive phyllorhizas 
remain distinct but. as development goes on, they become more 

86 Scientific Intelligence. 

crowded and, where they are joined together, a stem gradually 
makes its appearance upon which the leaves seem to arise second- 
arily. Although the described conditions are clear in these lower 
aquatic types, the development in spermatophytes and especially 
in those of terrestrial habit is so condensed that the phyllorhizas 
can be demonstrated only with difficulty. In support of his 
theory Chauveaud secures evidence from the gross morphology 
of the plants discussed and also from their minute anatomy. 

a. w. E. 

5. The Vegetation of Neiv Zealand (Engler und Drude's 
"Vegetation der Ercle," vol. 14) ; by L. Cockayne. Pp. xxii, 
364, with 65 plates, 13 text-figures and 2 maps. Leipzig, 1921 
(Wilhelm Engelmann). — Botanical investigations in New Zea- 
land date back to 1769 and the unique characteristics of the flora 
of this island* have been made known through a long series of 
works by many authors. Most of these investigations, however, 
have been along tax-onomic lines and it is only within the last 
twenty years that serious attention has been given to the study 
of vegetation — of plants considered en masse rather than as indi- 
viduals or species. With studies along this line is inseparably 
linked the name of Cockayne, and the present volume (in Eng- 
lish) embodies a clear and comprehensive summary of the work 
carried on by this author and his colleagues since about 1900, 
together with that contained in the scattered contributions of 
earlier writers. The greater part of the book is devoted to the 
"Vegetation of primitive New Zealand," but there are also 
sections on the physical geography and climate, the flora and its 
distribution, and the history of the flora. g. e. nichols. 

6. Les Mouvements des Vegetaux. — Du reveil et du sommiel 
des plantes; by Rene Dutrochet. Pp. vii, 121, with 21 text-fig- 
ures. Paris, 1921 (Gautier-Villars et Cie.,Editeurs). — Dutrochet 's 
essays on plant movements, which are here reprinted, were pub- 
lished in 1837 and exerted a profound influence on the develop- 
ment of plant physiology. The author was able to demonstrate 
the fact that many of these movements could be explained from 
a mechanical point of view and emphasized the important part 
played by osmotic phenomena in bringing them about. The 
present volume is the first botanical number of a series entitled, 
' l Les Maitres de la Pensee Scientifique. ? ' a. w. e. 

7. Die Pflanzemvelt Afrikas, inshesondere seiner tropischen 
Geoiete: Die dicotyledonen Angiospermen (Engler & Drude's 
"Vegetation der Erde," vol. 9) ; by A. Engler. Pp. vii, 878, 
with 338 text-figures. Leipzig, 1921 (Wilhelm Engelmann). — 
In this second part of the third volume of the Vegetation of 
Africa (the ninth of the series of plant geographical monographs 
edited by Engler and Drude) Professor Engler continues the 
discussion of the families of African plants commenced in Vol- 

Natural History. 87 

ume 2. The present volume completes the dicotyledonous Angio- 
sperms, extending from the Euphorbiaceae through the Corna- 
ceae. Engler was assisted in the preparation of this work by 
several of his friends and students, notably. Dr. Pax of Breslau, 
who is responsible for the exhaustive treatment of the Euphor- 
biaceae. Other collaborators include Drs. Brehmer, Brandt, 
Burret, Diehls, Gilg, Harms, Krause, Loesener, Mildbraed, Radl- 
kofer and Ulbrich. The method of treatment follows that of the 
earlier volumes and is essentially similar to that of the familiar 
Pflanzenreich. The discussion of the families is followed by a 
summary of the geographical relations of the African flora with 
examples of the various floristic elements of which it is com- 
posed ; and also by an account of the morphology, taxonomy, dis- 
tribution and origin of the characteristic xerophytes of the 
country. a. f. hill. 

8. Precis de Physiologic Vegetale; by L. Maquenne. Pp. 
175, with 4 text-figures. Paris, 1922 (Payot & Cie.). — Professor 
Maquenne 's work gives in condensed form the essentials of plant 
physiology. It represents a resume of a course of lectures 
given for many years at the Natural History Museum in Paris. 
The first chapter deals with certain physical and chemical 
phenomena which have an immediate bearing on the plant's 
activities. In the second chapter germination is discussed; in 
the third, fourth and fifth, the anabolic processes of the plant ; 
in the sixth, the respiratory processes; and in the seventh and 
eighth, the movements of water and other substances in the plant, 
together with the changes associated with maturity. The ninth 
and last chapter discusses the more important chemical sub- 
stances found in plants. a. w. e. 

9. The North American Slime-moulds: A descriptive list of 
all species of Myxomyeetes hitherto reported from the continent 
of North America with notes on some extra-limital species ; new 
and revised edition; by Thomas H. Macbride. Pp. xvii, 299, 
with 23 plates. New York, 1922 (The Macmillan Company). — 
The first edition of this valuable manual was published in 1899 
and contained 231 pages and 18 plates. The increase in size is 
largely due to the addition of 53 species to the flora of North 
America, including 12 proposed as new. The extra-limital 
species described number 13 and occur in various parts of the 
world. The most important change in the systematic portion 
of the work is the recognition of both orders and families, these 
groups being designated by the usual suffixes. In the first edition 
the orders had the family suffix while the families had the tribal 
suffix. The first 18 plates in the new edition are in large part 
redrawn from those in the first edition, while the additional 
plates illustrate the species since accredited to our flora. A com- 
parison of the two editions demonstrates the continued activity of 

88 Scientific Intelligence. 

North American students of the slime-moulds and the important 
results achieved during the last twenty-five years. a. w. e. 

10. Soil Conditions and Plant Growth; by E. J. Russell. 
Pp. xii, 405, with 32 text-figures. Fourth edition. London, 
1921 (Longmans, Green & Co.). — Originally published as one of 
a series of monographs on biochemistry, this book in its present 
form appears as the first of the "Rothamsted monographs on 
agricultural science." Presenting as it does a comprehensive 
survey of the physical, biological and chemical conditions of the 
soil as related to plant growth, a survey which three thorough 
revisions have kept well abreast with advances in scientific knowl- 
edge in this complex field, its value to students both in this and 
in related fields has been recognized from the outset. The fourth 
edition, more than half as large again as the first, retains the 
general character of the earlier editions, dealing broadly with the 
whole subject. g. e. nichols. 

11. A Handbook of the British Lichens; by Annie Lorrain 
Smith. Pp. 158, with 90 text figures. London, 1921 (British 
Museum). — This Handbook is based on the author's recent Mono- 
graph of the British Lichens, a much more extensive work in two 
volumes published by the British Museum. After a brief intro- 
duction in which the morphology of lichens is described, together 
with notes on their physiology, ecology, economic uses and phylo- 
geny, the author gives full descriptions of the families and 
genera of British Lichens, and makes it possible to determine the 
species by means of keys. Characteristic species of the more 
important genera are illustrated. Since many of the British 
species occur also in North America the Handbook ought to prove 
serviceable to students of the lichens on both sides of the Atlantic. 

a. w. E. 

IV. Miscellaneous Scientific Intelligence, 

1. The Outline of Science: A plain story simply told; edited 
by J. Arthur Thompson. Four volumes, with 40 colored plates 
and 800 other illustrations. New York and London, 1922 (G-. P. 
Putnam's Sons). — With the high degree of specialization that 
has taken place in recent years among the workers in the various 
fields of science there are now but very few who have been able to 
keep in touch with the marvelous discoveries of recent years and 
with the modern points of view in all these branches. Conse- 
quently there are but few whose knowledge of science is so 
far-reaching as to give them a broad view of the interrelations of 
the entire field ; and only those could hope to succeed in present- 
ing to the general reader a comprehensive story of science. The 
editor of these volumes stands pre-eminent in this class. 

The first volume, which is the only one of the series which has 

Miscellaneous Scientific Intelligence. 89 

yet appeared, gives bright promise of the phenomenal success of 
the whole. It consists of eight chapters, covering 296 pages, with 
more than two hundred illustrations and ten colored plates. 

The book begins with a survey of the universe, with a simply- 
stated but thrilling ' ' romance of the heavens. ' ' Then follows the 
story of evolution, with an account of the evolution of the earth, 
the hypothetical origin of life, the beginnings of life, and its 
evolution through the ages. The reader is next shown the mar- 
velous adaptations of organisms to their environment, the endless 
struggle for existence, followed by the ascent of man. Now 
comes an analysis of the evolutionary changes taking place in 
man and other organisms at the present day and the wonderful 
progress which may result from man's intelligent propagation 
of desirable types selected from the multitude of new forms 
which nature is constantly producing. The dawn of mind is 
next taken up for discussion and the reader is led from the simple 
reaction systems of the lower forms of life to the eventual devel- 
opment of intelligence and reason. 

With the completion of the storey of evolutionary biology the 
reader's attention is directed to that most fascinating and 
speculative subject, the foundations of the universe. In this 
final chapter of the book the atom is analyzed in terms that any 
one may understand, the electron is pictured, and the inferences 
as to the relation of matter and energy are pointed out. 

At the end of each article is a short list of the books in which 
the reader can follow the subject to which the chapter has been 
so auspicious an introduction. 

It will be readily admitted that the aim of the book "to give 
the reader a clear and concise view of the essentials of present- 
day science' 7 will be fully realized if the remaining volumes are 
of the character of the first w. r. c. 

2. Publications of the Smithsonian Institution, Washington, 
D. C; Charles D. \Valcott, Secretary. — The Annual Report 
for 1920 has been recently received. This is a volume of over 
700 pages, containing the Secretary's report already noticed (vol. 
1, pp. 95, 96) : also the General Appendix (pp. 145-690) embrac- 
ing twenty-seven papers of wide interest all illustrated by excel- 
lent plates. 

Explorations and Field Work in 1921. Pp. 128. — The paper 
(pp. 1-22) by the Secretary, giving a Summary of his continued 
work in the Canadian Rockies for 1921, opens this pamphlet and 
aside from its scientific value will charm the reader by the repro- 
ductions of the beautiful photographs taken by Dr. and Mrs. 
Walcott in the region studied to the northwest of the Lake Louise 

Other papers, by various authors in different regions, deal with 
paleontological. astrophysical, botanical, entomological and 

90 Scientific Intelligence. 

archeological field work. The admirable character of the illus- 
trations leaves nothing to be desired. 

3. Banking, Principles and Practice; by Ray B. Wester- 
field. — An important and comprehensive work in five volumes 
on the historical, legal, practical aild theoretical aspects of bank- 
ing has recently been published by the Ronald Press Company 
(New York City). The author, R. B. Westerfield, is assistant 
professor of political economy at Yale University and secretary- 
treasurer of the American Economic Association. Prof. Wester- 
field writes authoritatively from long experience and exhaustive 
study, and although the work is thoroughly up-to-date, embrac- 
ing the most recent developments, a conservative approach to the 
issues in discussion has been preserved throughout. The exposi- 
tion of the subject matter is admirably clear, and the various 
phases of banking are presented with as much conciseness as their 
numerous ramifications allow. The author has apparently spared 
no pains or caution in making the work completely reliable, so 
that it constitutes a valuable and accurate symposium for refer- 
ence study. The careful and orderly arrangement of chapter 
headings and subjects is well adapted to convenience and handy 

The first volume (pp. 207) deals with the fundamental princi- 
ples of money, credit and banking; the second (pp. 208-507) 
with the banking system of the United States, including full dis- 
cussion, historical and descriptive, of the Federal Reserve and 
the Federal Farm Loan systems; the third (pp. 511-809) with 
domestic banking, covering cash and deposit operations, bank 
management, administration of departments, and other import- 
ant branches of the subject. The fourth volume (pp. 810-1080) 
continues the discussion of domestic banking with complete 
information concerning earning assets, and the fifth (pp. 1081- 
1370) gives all the details of the foreign division, comprising 
foreign exchange, the handling of foreign operations, collections, 
letters of credit, and miscellaneous departments. 

The book admirably fills the needs of more than one class of 
people ; First, the student who wishes to know what banking is 
and how it is done ; second, the employe who desires to perfect 
himself in his subject and secure promotion; third, the banker 
of experience who will use the book for reference and latest 
methods; fourth, the' teacher who could find no better text book 
in so wide a field ; . fifth, every inquiring mind. 

''Banking, Principles and Practice" provides a compendium 
of practical information which unites in a single work the many 
phases of an entire subject. As such, it fills a definite need and 
does so in a comprehensive and authoritative manner. 


4. Civic Science in the Home. Pp. 416. Civic Science in the 

Miscellaneous Scientific Intelligence. 91 

Community. Pp. 430; by George W. Hunter and Walter G. 
Whitman. New York, 1921 (Amer. Book Company). — These 
books cannot be criticized for narrow specialism. They compass 
the whole wide domain of General Science, and attempt to relate 
this science to citizenship. The books are intended for public 
elementary school use and their scope is defined by their authors 
in the following statement: "In short, Civic Science plans to 
lead the child in a manner which is both logical and psychologi- 
cal from the simple factors which make up his environment as a 
living thing to the complex combinations and interactions which 
have arisen through what we call civilization." 

Consequently the chapters range from flies, foods, and pests to 
street lighting, automobiles, plumbing, eugenics and euthenics. 
The task of organizing such eclectic material into pedagogical 
units is, of course, a difficult one. The authors have depended on 
the problem setting or project method to accomplish such unifica- 
tion. They have also made very free use of the blank score card 
as a teaching device. Thus we have score cards for the "home," 
"the natural resources of my environment," "water in my 
home," "foods in my home," "removal of wastes from my 
home," etc. 

The volumes are attractively printed, crowded with pictures 
and diagrams and references and, therefore, constitute a useful 
source book for the teacher. The proper teaching of elementary 
school science unfortunately depends to a very slight degree upon 
textbooks ; but, even without a good teacher, these books will 
awaken a popular interest in science and an appreciation of its 
everyday importance, thus anticipating in a way the "Science 
Service" of the American Association for the Advancement of 
Science. Arnold gesell. 

5. Memoirs of the Queensland Museum, vol. VII, parts II and 
III ; edited by the Director, Heber A. Longman. — Part II (pp. 
65-80, with 4 plates) contains an account by the Director of a new 
genus of fossil marsupials, Eury zygoma. This was obtained 
from the Post-Tertiary deposits in the Darling Downs. It is 
described as a remarkably bizarre monster of the Nototherium 
group {Eury zygoma dunense) with a skull, the width of which 
exceeds the length by 46 mm. 

Part III (pp. 81-240, plates VIII to XII) contains an article 
by T. D. A. Cockerell of Colorado on Australian bees, with a 
catalogue by Henry Hacker ; also the second of the papers on 
Queensland fishes by A. R. McCulloch ; a new Nyctimene by 
H. A. Longman and on Coleoptera, mostly from Queensland, by 
A. M. Lea. 

6. United States Life Tables, 1890, 1901, 1910 and 1901-1910 ; 
prepared by James \V. Glover. Pp. 496, quarto. AVashington, 
1921. (Bureau of the Census, Samuel L. Rogers, Director.) — 

92 Scientific Intelligence. 

This volume, in addition to the full explanatory text, mathemati- 
cal theory, computations, graphs and original statistics, gives also 
tables of United States life annuities, life tables for foreign coun- 
tries and mortality tables of life insurance companies. The 
scope of the work is, therefore, as wide as the subject itself; 
Mr. Glover, who has had charge of it, is not only the expert 
special agent of the Census Bureau but also professor of mathe- 
matics and insurance in the University of Michigan. 

The tables show the rates of mortality and expectation of life 
for all classes of people residing in the country. It is interesting 
to note that the two decades from 1890 to 1910 show an improve- 
ment in mortality conditions for men and women under 50 years, 
while above that age conditions were stationary. Further women 
rank much higher than men, persons living in the country higher 
than those residing in cities, whites than negroes, and for most 
ages the native born than those of foreign birth. 

The work is divided into eight parts, the first five being 
designed for the general public, the remainder for the specialist. 
Part I (pp. 23-49) carefully read will give the average reader 
all the information ordinarily required, and enable him to con- 
sult the many tables with intelligence. 

7. Public Opinion; by "Walter Lippmann. Pp. x, 427. New 
York, 1922 (Harcourt,- Brace & Howe). — Mr. Lippmann 's 
book falls into two parts, nearly equal in size. The first part is 
mainly a study in psychology, as is indicated by its headings : 
The world outside and the pictures in our heads, approaches to 
the world outside, stereotypes, interests. The author, who was 
graduated from Harvard in 1910, has kept alive the interests 
developed under AVilliam James and others there, and applies his 
analysis to the conditions, both individual and social, underlying 
the formation of public opinion. In the latter part of the book 
he follows his subject into the field of politics, taking as his main 
topics : The making of a common will, The image of democracy, 
Newspapers, Organized intelligence. 

The author has observed the action of the modern political sys- 
tem, both from the outside as journalist, and from the inside as 
a worker in practical politics, and, during the latter part of the 
war, in the service of the government. He illustrates his mean- 
ing with concrete examples from recent history, and so does much 
to help the reader over the hard places in doctrine. The book is 
interesting. It is, furthermore, important. The author makes a 
determined effort to see things as they are, not as they have been 
supposed to be. If he has not the experience of political sages 
like Morley and Bryce he has a freedom from tradition, a fresh- 
ness of imagination, and a vigor of attack which make his analy- 
sis of present conditions well worth attention. He does not offer 
a panacea for the ills of the time, but in his concluding chapters 
he discusses in a sober and practical way some measures of 
reform. His proposal, in brief, looks to a system in which 

Miscellaneous Scientific Intelligence. 93 

experts, trained and organized for the purpose, shall be set to 
ascertain facts of public interest, and so provide at least a firm 
basis for public opinion. olive day. 

8. Publications of the Carnegie Foundation for the Advance- 
ment of Teaching (522 Fifth Avenue. New York Gity). — The 
following publications of the Carnegie Institution have been 
recently issued. (See earlier, vol. 3, pp. 157, 307.) Bulletin 
XVI. Education in the Maritime Provinces of Canada. Pp. 
vii. 50, with a map (frontispiece). — Dr. AY. S. Learned of the 
Carnegie Foundation and Dr. Kenneth C. M. Sells of Bowdoin 
College have prepared this study of the present provincial situa- 
tion in higher education in Nova Scotia and New Brunswick, 
discussing also the elementary and secondary school conditions. 
The conclusion reached is the desirability of the union of six 
small universities in the Canadian provinces of Nova Scotia and 
New Brunswick to make one strong university at Halifax. The 
institutions concerned are King's College at Windsor, N. S., Dal- 
housie University at Halifax, Acadia University at AYolfville, 
X. S.. Mt. Allison University at Saekville. X. B.. St. Francis 
Xavier's University at Antigonish, X". S., and the University of 
New Brunswick at Fredericton. X. B. It is proposed to raise 
S4. 500.000 to accomplish the purpose proposed. 

The plan suggested is an adaptation of English collegiate 
organization. Each college would maintain its own student resi- 
dence, class-rooms, chapel, etc.. where most of the freshman and 
sophomore work would be conducted, while advanced courses, all 
honor course*, and the expensive laboratory sciences would be 
taught in a central university controlled by a joint board and sup- 
ported in part by the provinces. This arrangement would make 
possible the equipment of a first-class university, while sacrificing 
few or none of the advantages of the small colleges. It would 
permit the denominational college with small endowment to com- 
mand the most extensive university privileges and at the same 
time retain and strengthen its denominational character. An 
attractive feature of the scheme from an American point of view 
is the method suggested for the better organization of large 
bodies of students — an acute problem in American colleges. 

Other publications are : 

No. 215-A. History of Domestic and Foreign Commerce of 
the United States: by E. K. Johxsox, T. AY. Van Metre, G. G. 
Heebxer, and D. S. Haxchett. with an introductory note by 
H. AY. Parnam. Pp. xxiv. 7(31, 10 maps. A second "photogra- 
phic edition ($3.00). 

No. 309. Gaseous Exchange and Physiological Requirements 
for Level and Grade Walking* by Hexry Moxmouth Smith. 
Pp. viii. 310: 1 plate. 42 figs. ($6.00). 

No. 316. Development and Activities of the Boots of Crop 
Plants : A Study in Crop Ecology ; by J. E. AYeaver. F. C. Jean 
and J. AY. Crist. Pp. vi. 116. 14 pis.. 42 figs. ($2.00). 

94 Scientific Intelligence. 

No. 320. Genetic Studies of Rabbits and Rats; by W. E. 
Castle. Pp. 55, 2 pis., 7 figs. ($1.00). 

9. American Association for the Advancement of Science. — 
The seventy-fifth meeting of the Association was held in Salt Lake 
City on June 22 to 24, under the auspices of the Pacific Division, 
of which Dr. Barton \V. Evermann is president. 

10. Observatory Publications. — The University of Cincinnati, 
Jermain G. Porter Director, has recently issued No. 19 of its 
Publications. This gives a Catalogue of 4683 Stars for the 
Epoch 1900, observed by Eliott Smith and prepared for publi- 
cation by the director. It presents the positions of all stars 
determined with the meridian circle from January, 1907, to 
December, 1921. The proper motions of the greater part of the 
stars were computed in No. 18 ; other stars are those of the Boss 
preliminary general catalogue. Computations have been made 
by the Director assisted by Dr. Yowell. It is announced that 
the next meridian work published by the Observatory will be 
reduced to 1925. 

The observatory of Krakow has issued an eight-page publi- 
cation giving the minimos of Algol and RW Tauri in 1922, 
reduced to the meridian of Greenwich. 

The Annual Report of the U. S. Naval Observatory at "Wash- 
ington has also been received. 

The National University of La Plata has issued part 2 of vol- 
ume VI, by Numa Tapica, giving micrometric measurements of 
double and vicinal stars. From the same source comes a deter- 
mination of the orbit of the planetoid (796) Saria by Hugo A. 

The Hector Observatory at Wellington, New Zealand, pub- 
lished some time since a reprint (Bulletin 33) on observations of 
Southern variable stars by C. E. Adams; also a brief method of 
calculating occultations of stars by the moon by C. J. Westlaxd. 

The observatory at Lyons, France, publishes a Bulletin of 
which No. 4 for April, 1922 (4th year) has been received (H. 
Georg, Editeur). The annual subscription is only 12 francs. 


Professor George Simonds, the English botanist, died on May 
4 at the age of sixty-nine years. 

Arthur Bacot, entomologist to the Lister Institute, died at 
Cairo on April 12 as the result of infection contracted when 
prosecuting investigations on the etiology of typhus. 

Louis Antoine Ranvier, the veteran French histologist, died 
recently at the age of eighty-seven years. 

Dr. Henry Marion Howe, the distinguished metallurgist, 
since 1897 professor in Columbia University, died recently at 
the age of seventy-four years. 

Ward's Natural Science Establishment 

A Supply-House for Scientific Material. 

Founded 1862. Incorporated 1890. 

A few of our recent circulars in the various 
departments : 

Geology: J-32. Descriptive Catalogue of a Petrographic Col- 
lection of American Eocks. J- 188 arid supplement. 
Price-List of Rocks. 

Mineralogy : J-220. Collections. J-238. Minerals by Weight. 
J-224. Autumnal Announcements. 

Paleontology: J-201. Evolution of the Horse. J-199. Pale- 
ozoic index fossils. J-115. Collections of Fossils. 

En tomology : J-33. Supplies. J-229. Life Histories. J-230. 
Live Pupae. 

Zoology: J-223. Material for dissection. J-207. Dissections 
of Typical Animals, etc. J-38. Models. 

Microscope Slides: J-189. Slides of Parasites. J-29. Cata- 
logue of Slides. 

Taxidermy: J-22. North American Birdskins. Z-31. General 

Human Anatomy: J-37. Skeletons & Models. 

General : J-228. List of Circulars & Catalogues. 

Ward's Natural Science Establishment 

84-102 College Ave., Rochester, N. Y., U. S. A. 


number consisting of 100 to 120 pages). Editor : EUGENIO RIGNANO. 

This is the only review which has a really international collaboration ; which is 
of world-wide circulation and occupies itself with the synthesis and unification of 
knowledge, in the history of the sciences, mathematics, astronomy, geology, physics, 
chemistry, biology, psychology and sociology. 

It is the only review which, by inquiries among eminent scientists and writers 
(on: The philosophical principles of the sciences; Fundamental astronomical and physical 
questions of current interest; Contributions given by the various countries to different 
branches of knowledge; Question of vitalism; Social question; International questions 
raised by the world war), makes a study of the problems interesting scientific and intel- 
lectual circles throughout the world. 

It has published articles by Messrs. : Abbot -Arrhenius =Ashley- Bayliss-Beichman- 
Bigourdan - Bohlin - Bohn = Bonnesen - Borel - Boat y = Bragg - Bruni = Burdick - Carver - 
Cauller y - Chamberlin - Charller = Claparede - Clark - Costantin - Crommelln - Crowter - 
Darwin - Delage - De Vrles •= Durkheim =■ Eddington - Edgeworth = Emery = Enrlques - 
Fabry - Findlay - Fisher - Fowler = Golgi = Gregory - Harper - Hartog = Helberg - Minks - 
Hopklns-Inigues-Innes-Janet-Kaptein-Kaye=Kidd-Langevin<=Lebedew°Lloyd Morgan - 
Lodge - Loisy - Lorentz - Loria =Lo we 11 - MacBride - Meillet -Moret-M uir=Peano -Picard - 
Polncare - Palseux = Rabaud - Rey Pastor - Righ i = Rignano= Russell=Rutherford=Sagnac - 
Sarton - Schiaparelll- Scott - See - Sherrington = Soddy - Starling- Svedberg- Thomson • 
Thornd ik» -Turner -Volterra -Webb -Weiss = Zeeman and more than a hundred others. 

" SC1ENTIA" publishes its articles in the language of its authors, and joins to the 
principal text a supplement containing the French translations of all the articles that 
are not in French. (Write for a Specimen Xumber to the General Secretary of 
"Scientia", Milan, sending, - to defray postal and other expenses,- 2 Francs in 
stamps of your country.) Annual subscription: 40sh.,or 10 dollars post free. 

Orhce : 43 Foro Bonaparte, Milan, Italy 

Publishers: WILLIAMS & NORG ATE - London ; FELIX ALCAN- Paris 




Art. I. — The Melting of Potash Feldspar; by G. W. Mokey 

and N. L. Bowen, 

Art. II; — Triassic Reptilian Order Thecodontia, by F. 

ton Htjene, , 22 

Art. III. — A Discussion of Triple Salts; by H. L. Wells, 27 
Art. IV. — Horned Eocene Ungulates; by E. L. Troxell, . . 31 
Art. Y. — The Genus Hyrachyus and its Subgroups; by E, 

L. Troxell (with Plate I), 38 

Art. VI. — A New Occurrence of Ilsemannite ; by C. W. Cook, 50 
Art. VII. — On the Zonal Division and Correlation of the 
Silurian of Bohemia ; by J. Perner, with the collaboration 
of O. Kodym, 53 


Chemistry and Physics. — A New Process for the Industrial Production of Barium 
Hydroxide for Treatment of Molasses in Sugar Penning, Deguide and Bode: 
Advanced Course of Instruction in Chemical Principles, A. A. Noyes and 
M. S. Sherrill, 78. — Determination of Sulphur in Iron and Steel. H. B. 
Pulsifer: Organic Chemistry, V. v. Richter, 74. — Friction and Lubri- 
cation, Hardy and Doubleday, 75. — Power Alcohol, G. W. Monier- 
Williams, 76. — The Journal of Scientific Instruments: La Theorie Ein- 
steinienne de la Gravitation, 77. 

Geology and Mineralogy. — Gravity Anomalies and their Geological Interpre- 
tation, 78. — Publications of the United States Geological Survey, G. O. 
Smith, 79. — Die Eruptivgesteine des Kristianiagebietes IV; Das Fengebiet 
in Telemark, Norwegen, W. C. Brogger, 80. —Mineral Resources of the 
Philippine Islands for 1919 and 1920, 82. -A List of new Crystal Forms 
of Minerals: Handbook and Descriptive Catalogue of Gems and Precious 
Stones in the U. S. National Museum, G. P. Merrill, etc. : Virginia Geo- 
logical Survey, T. L. Watson: The Topographic and Geological Survey 
of Pennsylvania, G. A. Ashley: Geology of Drumheller Coal Field, Alberta, 
J. A. Allan, 83. — Potash in a new area of Texas, 84. 

Natural History. — Arctic Alcyonaria and Actinaria, A. E. Verrill: Genetics, 
An Introduction to the Study of Heredity, H. E. Walter, 84. — A Natur- 
alist in the Great Lakes Region, E. R. Downing: La Constitution des 
plantes vasculaires revelee par leur Ontogenie. G. Chauveaud, 85. — The 
Vegetation of New Zealand, L. Cocayne: Les Mouvements des Vegetaux, 
R. Dutrochet: Die PfLanzenwelt Afrikas, insbesondere seiner tropischen 
Gebiete, A. Engler, 86. — Precis de Physiologic Vegetale, L. Maquenne: 
The North American Slime-moulds, T. H. Macbride, 87. — Soil Conditions 
and Plant Growth, E. J. Russell: A Handbook of the British Lichens, 
A. L. Smith, 88. 

Miscellaneous Scientific Intelligence. — The Outline of Science, J. A. Thompson, 
88. — Publications of the Smithsonian Institution, C. D. Walcott, 89. — 
Banking, Principles and Practice, R. B. Westerfield: Civic Science in 
the Home, G. W. Hunter and W. G. Whitman, 90. — Memoirs of the 
Queensland Museum: United States Life Tables, J. W. Glover, 91. — Public- 
Opinion, W. L-ippmann, 92. — Publications of the Carnegie Foundation for 
the Advancement of Teaching, 93. — American Association for the Advance- 
ment of Science: Observatory Publications, 94. 

Obituary.— G. Simonds: A. Bacot: L. A. Ranvier: H. M. Howe, 94. 

Library, U. S. Nat. Museum. 


Established by BENJAMIN SILLIMAN in 1818. 




Editoe: EDWARD S. DANA. 



of Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 

of Washington. 



No. 20— AUGUST, 1922. 


19 22. 


sr^ • 

Published monthly. Six dollars per year, in advance. $6.40 to countries in the 
Postal Union ; $6.25 to Canada. Single numbers 50 cen^s ; No. 271, one dollar. 

Entered as second-class matter at the Post Office at New Haven, Conn., under the Act 
of March 3, 1879. 


By H. RlES, PH.D., and THOMAS L. WATSON, Ph.D. 

Elements of Engineering Geology 

More than a condensation and simplification of the authors' larger text 
'Engineering Geology, 1 since it has involved complete rewriting of many parts 
of the larger book and the amplification of other parts. 

365 pages. 5^ by 8*4. 252 figures. Cloth, $3.75. 

Engineering Geology 

Outlines those fundamental principles which relate to engineering practice 
including the character of rocks, their use in building, their structure and the 
geological conditions which affect and control underground waters. 
722 pages. 6 by 9. 249 figures, 104 plates. Cloth, $5.00. 

By H. RlES, PH.D. 

Clays: Their Occurrence, Properties and Uses 

Describes the occurrence, properties, methods of mining and the manu- 
facturing and uses of a vast variety of clays. 

554 pages. 6 by 9. 112 figures, 44 plates. Cloth, $5.00. 

Economic Geology 

Presents, first, the non-metallic minerals and then the metallic, thus lead- 
ing the reader from the simpler to the more complex phases of the subject. 
856 pages. 6 by 9. 291 figures, 75 plates. Cloth, $5.00. 


History of the Clay- Working Industry in the 
United States. 

Mentions the important deposits of clay and takes up the developments in 
manufacturing methods. 

270 pages. 6 by 9. Illustrated. Cloth, $2.50. 

Send for any or all of these books on Free Examination terms. 



432 Fourth Avenue 
New York 

Montreal, Quebec 
Renouf Publishing Company 
AJS 8.22 



Art. VIII. — Colloids in Geologic Problems; by George 

D. Hubbard. 1 

Introduction. — It lias long been recognized that solid 
matter occurs in more than two states. The terms 
"crystalline" and "amorphous" exclude quite a large 
body of materials, some of which are organic and some 
inorganic. The other condition has been called the col- 
loidal state. Many substances which occur crystalline 
can be prepared in the colloidal state without any com- 
positional difference. Amorphous substances were in 
many instances once in colloidal condition. Perhaps all 
substances can ultimately be prepared in this state 
whether they are normally crystalline or amorphous. 

In the case of some substances, there seems to be no 
line fixed between the crystalline and the colloidal state. 
In the same way, there seems to be no line which can be 
definitely drawn between a true solution and a colloidal 
solution. In practice we say a substance is in the col- 
loidal state when it will not dialyze or diffuse through 
certain membranes, as egg skin, bladder, goldbeater's 
skin, or parchment paper. Molecular dispersions are in 
true solution; dispersions of much larger aggregates (a 
hundred, more or less, of molecules in one particle) are 
in colloidal solution. Such particles are too large to pass 
through the membrane. 

Bancroft 2 says, "Colloid chemistry is the chemistry of 
grains, drops, bubbles, filaments, and films." Grains of 
solid particles, drops of fluids, bubbles of gas, filaments 
very small in two directions, films small only in one direc- 

1 The author is deeply indebted at many points in the paper to Dr. Harry 
X. Holmes of Oberlin College for suggestions and criticism. His General 
Chemistry and Colloid Manual have been of much help also. This paper was 
presented to the Geologic Section of the Ohio Academy of Science, April 15, 

2 Bancroft, W. D., Applied Colloid Chemistry, p. 2 (McGraw-Hill Co.). 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 20.— August, 1922. 

96 G. D. Hubbard — Colloids in Geologic Problems. 

tion, — these constitute the field of the student of colloids. 
The smallest particle visible with the aid of the best com- 
pound microscope is about 100 millimicrons in diameter, 
while the largest molecules approach a diameter of 
1 millimicron. Materials in the colloidal state have parti- 
cles ranging between these limits. 3 

The term "colloid" or "colloidal state" has been 
expanded to include not only solid gels, but suspensions 
and emulsions. The former may be solids in liquids, as 
the very fine sediments in roily water, liquids in liquids, 
or even solids in solids, as the blue rock salt which is a 
suspension of finely divided metallic sodium in crystalline 
halite. In suspensoids the solid particles do not power- 
fully absorb or hold large quantities of water. In emul- 
soids, on the other hand, the particles very powerfully 
hold or absorb large quantities of water, forming gels. 
Such gels have been made solid, containing as high as 
99.87% water. 

Gels are substances in the colloidal state which have 
set or become apparently solid, even though they contain 
a large percentage of water. The fruit juice (acid), 
water, and sugar in the jelly cup have set because of a 
small percentage of pectin. All are essential to the 
phenomenon. Certain minerals, for example most ortho- 
silicates and some zeolites, on some evaporation after 
solution in hydrochloric or nitric acid, will set in beautiful 
gels because of the silicic acid liberated. Sodium silicate 
or water glass sets in the same way when treated with 
acetic or some other acid, which makes a salt with the 
sodium and leaves the silica to gel. 

As early as 1861 Thomas Graham 4 knew that dissolved 
crystalline substances diffused in and out of silica gels, 
and that reactions occurred between salts in solution in 
the gels and solutions from outside which diffused in. 
The famous Liesegang rings were discovered in 1896. 
Many of these were rings of metallic crystals or crystal- 
line salts, formed in silica gels by reactions between two 
solutions, one of which was in the gel and the other was 
gradually diffusing through the gel. The well-developed 
crystal had time to form because the solutions came in 

3 Holmes, H. N"., Colloid Chemistry, 1922 (Wiley and Sons). 
4 Hatschek, E., and Simon, A. L. Trans. Inst. Mining and Metallurgy, Lon- 
don, 21, 451-480, 1911-12. 

G. D. Hubbard — Colloids in Geologic Problems. 97 

contact so slowly. Bechhold in 1905, Ziegler in 1906, and 
Hatschek in 1911 continued experiments with rings 
in gels. 

Mineral Genesis through Colloidal State. — Liesegang 
suggested after his studies on the rings in gels that the 
banding of agates might be due, in many cases, to the 
slow diffusion of iron salts through silicic acid gels which 
had previously been laid in cavities or crevices. He also 
suggested that this process might explain the formation 
of large crystals in or on quartz in places where igneous 
activity had not operated, and where they therefore could 
not be due to its heat and gases. 

In the discussion at the Academy meeting, two points 
were made. Prof. Gr. F. Lamb reported orally the finding 
of good agates in recent clays ; the data seem to be want- 
ing as to whether they developed in preexisting cavities 
or made their own spaces. Agates in the major cavities 
of buried bones testify to the recency also of agate- 
making, and as well to the rate of making. Long ages 
are not necessary. No evidence was presented as to 
whether or not the agates came through the gel stage, or 
how the banding originated. 

As early as 1884, Clarke tells us, Bourgeois 5 made opal 
synthetically from gels of Si0 2 . Clarke 6 states also that 
the gelatinous silica formed by the solution of silicates 
in mineral acids becomes, upon drying, an amorphous 
mass essentially identical with opal. He also reports 
that Schafhautl heated a solution of colloidal silica in a 
Papin digester and obtained a crystalline deposit of 
quartz ; while de Senarmont heated gelatinous silica with 
water and carbonic acid gas to temperatures between 200° 
and 300 C and obtained a crystalline quartz. Chrust- 
schoff, 1873, obtained quartz from an aqueous solution of 
colloidal silica by heating to 250° C. for several months. 
The heat apparently hastened the process, which would 
have reached the same end in a longer time without so 
much heat. Silica is dissolved when silicates decompose, 
and is redeposited by evaporation as opal if still in an 
acid solution, but when alkalies are present it crystallizes. 

When the Simplon tunnel in central Switzerland was 
put through the Alps, the engineers found a vein of silica 

'° Bourgeois, L., Production of Minerals Artificially, p. 93. French. 
G Clarke, F. W., Data of Geochemistry, U. S. G. S. Bull. 616, p. 357. 

98 G. D. Hubbard — Colloids in Geologic Problems. 

in the gel condition or colloidal state, thus showing that 
the silica gel is not purely an artificial product. If the 
ancient scientists who gave the name "quartz" to rock 
crystal had found this vein and watched it through its 
transformation to crystalline quartz, they might have had 
some ground for their theory that it was made of water 
which had become so thoroughly frozen in a very cold 
winter that its solid state had become permanent. 

The_experiments of Liesegang, and of many other col- 
loid chemists who have followed him, suggest a method 
for the formation of the quartz-gold vein. They have 
shown that if a gold solution, e.g. AuCl 3 , is mingled with 
a silica gel when the latter sets, and then is treated with 
a reducing agent diffusing through the gel, crystals of 
gold will form in the gel. Oxalic acid placed on top may 
serve as this reducing agent, H 2 S in water, S0 2 , or CO 
will serve the same purpose. Sodium sulphite, and 
especially FeS0 4 , do occur in solutions in nature and 
would be suitable reducing agents there. 

Holmes 7 has shown that a preparation of silicic acid gel 
made 0.1N with respect to potassium iodide, and then 
covered with 0.5N mercuric chloride, gives in the gel after 
a few days bands of red crystalline mercuric iodide. In 
some cases as many as 40 rather sharply marked bands 
occurred in a distance of 8 cm. In a similar manner, 
bands of copper chromate, cuprous oxide, basic lead 
iodide, basic mercuric chloride, and other substances were 
formed. Bands of colloidal gold with scattered crystals 
of gold developed in a preparation of gold chloride and 
H 2 S0 4 with oxalic acid above the gel. 

If the silica gel were in one crevice and an intersecting 
crevice should be conveying one of these reducing solu- 
tions, the conditions of the laboratory would be very 
nearly duplicated. The only difference would be that the 
solution need not become continually more dilute, but 
might remain essentially the same in composition for 
years. In such case there might be no occasion for the 
banding of the salts, or if bands developed, they would 
no doubt be evenly spaced. Banding is explained 8 as a 
result of a reduction in concentration of the solution in 
the gel just in front of the ring, so that the diffusing 

7 Holmes, H. N., Jour. Amer. Chem. Soc, 40, 1187-1195, 1918. 
"' Holmes, ibid. 

G. D. Hubbard — Colloids in Geologic Problems. 09 

solution must needs go beyond the clear space to find 
sufficient concentration for the development of another 
band. If the flowing solution remain constant in quantity 
and condition, would not the bands maintain uniform 
spacing instead of becoming farther and farther apart as 
in the laboratory experiment? 

The laboratory experiment also throws light on the 
formation of metallic copper in quartz veins; The exper- 
imentor used a silica gel dialyzed with acetic acid I sul- 
phuric would have served the same purpose), and mixed 
with his gel, before it set. a solution of copper sulphate. 
The reducing agent was placed on top. and the tiny tetra- 
hedrons of metallic copper began to shine- in the gel. If 
the gel was free from air or any oxidizing agent, the 
crystals preserved their copper luster, but without due 
precautions, they oxidized dark to black. Clusters and 
chains of tetrahedrons formed as the process continued. 
After many months this process produced a silica gel 
vein with appreciable quantities of metallic copper dis- 
seminated through it. If one wishes to repeat or extend 
this work. Prof. Holmes 9 has given full directions for 
preparing silicic acid gels. 

In these cases, all that is necessary to make the typical 
mineralized quartz vein is to allow the gel to crystallize. 
a change which has already been explained as a laboratory 
experiment, and which seems perfectly possible as a 
natural geologic phenomenon. 

Lhnonite seems to furnish an illustration of an ore once 
in a colloidal state. There are six hydroxides of iron, 
made by combining one to six molecules of water with the 
ferric oxide, only one of which, goethite, is known in the 
crystalline form. This is the hydroxide with one mole- 
cule of water to one of Fe 2 3 . Clarke in the "Data of 
Geochemistry. 7 ' p. 531, quotes Van Bemmelen as saying 
that probably no formulas should be given to any of these 
ferric hydroxides, but rather they should be regarded as 
polymers of the first, or colloidal, complexes of two or 
more polymers. Van Bemmelen in a number of papers 
published between 1888 and 1904 so regards them, and 
Xicolardot and Spring both consider the several hydrox- 
ides polymers of the first. Thus the ferruginous slimes 
in bogs and around springs, and sometimes in stalactites 

& Holmes.. H. X.. Silicite Acid Gels. Jour. Phys. Cheni., 22, 510-519, 1915. 

100 G. D. Hubbard — Colloids in Geologic Problems. 

in caves, as at Mount Tabor or Beanie's Gave, Ohio, many 
of which are precipitated by organic agencies, are truly 
colloids which may in the course of time be dehydrated 
and pass over into the crystalline goethite, or even into 

Some bauxite was formed by passing through a col- 
loidal state. Colloidal alumina is soluble in water and 
has been carried some distance before coagulation has 
taken place. The coagulation is a result of the meeting 
with electrolytes in solution, whose charge is different 
from that of the colloid, and therefore the alumina loses 
its negative charge ; the solution first becomes turbid, 
then the hydroxide actually flocculates and collects into 
larger and larger lumps, and finally falls to the bottom. 
Apparently the continued growth of these lumps by the 
successive addition of layers of aluminum hydroxide 
builds up the oolitic and pisolitic bodies of typical bauxite. 
Some bauxite is certainly not of this origin. 

These gel ores of iron and aluminum start by decom- 
position of silicate rocks in the weather zone. Descend- 
ing waters carry them down into other rocks, or surface 
waters carry them away to marshes and bogs where the 
the precipitating conditions are found, and the ores are 
dropped. In the cooler, higher latitudes the iron ores 
are formed, while the alumina and silica remain 
together 10 ; and in the warmer or subtropical climates 
the laterites or bodies of bauxite are formed. Thus the 
ore deposit may witness to the character of the climate in 
force when the deposit was made. The precipitation in 
other rocks would probably take place above the water 
table; hence in high and dry regions there would be a 
long vertical range for deposition, which could take place 
anywhere between the surface and the water table. 

It has been pointed out by Krusch 11 that the manganese 
oxides may also be laid down as gel ores, and subse- 
quently become crystalline. He makes a point also of the 
fact that this colloidal ground mass of manganese oxides 
and hydroxides is capable of adsorbing other substances 
which may permeate it, e.g., barium and potassium com- 
pounds. In this way, he accounts for psilomelane, a 
manganese mineral containing barium and potassium. 

10 Schlosingj T., Compt. Bed., 132, p. 401, 1901. 

11 Krusch, P., Mining and Scientific Press, 170, 418, 1913. 

G. D. Hubbard — Colloids in Geologic Problems. 101 

F. W. Clarke in his "Data of Geochemistry," p. 533, 
accepts psilomelane, and some other less known man- 
ganese ores as colloidal complexes. In this connection 
he also mentions lampadite, a cupriferous hydrate of 
manganese. The copper may be present simply by 

The silicate nickel ores have always been a confusing 
group, but an interpretation involving colloidal sub- 
stances seems to simplify the matter. It will be remem- 
bered that these ores are hydrous magnesium nickel sili- 
cates. The non-metalliferous foundation seems to have 
accumulated first as a gel, 12 and subsequently to have 
adsorbed some nickel compound, perhaps nickel silicate, 
and a magnesium compound, and thus a mixture of two 
substances occurs whose proportions may vary greatly 
within short distances. 

Coagulation or Flocculation of Sediments. — It has long 
been known that even the finest sediments which come out 
to the sea in the waters of our great rivers fall to the 
bottom long before they get far into the sea, and that 
the main body of the ocean is made of comparatively 
clear water. This speedy precipitation of sediments has 
also been ascribed to the saltiness of the sea, but the con- 
nection had not been worked out until recent studies in 
colloids and their relations to electrolytes. 

An electrolyte is a salt, or in some. cases a base or an 
acid, in water solution through which electric current 
flows. Salts are generally better electrolytes than acids. 
The theory is that the electrolyte is ionized, e.g., sodium 
chloride becomes Na + and CI" in solution, magnesium 
chloride becomes Mg ++ and 2 CI" in solution. The higher 
the valence of the base part of the salt (the positive ion), 
the more powerful the electrolyte, as a coagulant of nega- 
tively charged suspensions, and the power seems to bear 
no definite relation to the valence. Aluminum chloride, 
for example, is hundreds of times more potent than 
sodium chloride. The charge on the electrolyte and on 
the colloid must be of opposite sign. The aluminum ion 
of A1 2 (S0 4 ) 3 has a positive charge, and aluminum is 
trivalent in the sulphate. Since most colloids are nega- 
tive, aluminum sulphate is one of the most effective floc- 

12 Kruseh, P., loc. cit. 

102 0. D. Hubbard — Colloids in Geologic Problems. 

These salts are all three in the sea. The sodium chlo- 
ride makes up fully three-fourths of all the salts there, 
while the alum constitutes a very small fraction of one 
per cent, but because of its greater activity, it may be 
as efficient as the much more abundant sodium chloride. 
For our purposes, however, it is not material which does 
the work. A much more important question is, how long 
have the seas been salt enough to coagulate the fine clays 
in colloidal suspension as they come down? "We have 
shale rocks made of such clays of all ages back through 
the Mesozoic, Paleozoic, and perhaps well into the Proter- 
ozoic. Such clays of course would settle out of quiet 
waters without the assistance of an electrolyte, but such 
waters certainly would be difficult to find on the conti- 
nental shelf. Therefore we may safely assume that these 
shales made of extremely fine clays were largely precip- 
itated by electrolytes, and that the sea was salt enough in 
those Proterozoic days to effect their coagulation. 

Experiments show that one percent, sodium chloride 
solution will coagulate clays rather quickly and that one- 
tenth per-cent. alum solutions will bring them down in a 
few hours, so we need not assume that the sea had a 
greater saltiness than, say, 25 per cent, of its present 
salinity in order to bring down the clays. In this connec- 
tion it may be interesting to note that the age of the 
ocean has been calculated on the basis of the NaCl in the 
water, assuming essentially the same rate of inflow as 
at present, and only trifling losses as compared with the 
present salt in the sea. The latter assumption is easily 
established. The former is probably safe. Dividing the 
rate by the amount of salt in the sea, Chamberlin and 
Salisbury arrive at the figure of 370,000,000 years. If 
anything like one-fourth the present saltiness would be 
necessary before effective precipitation should take place, 
then these old shales may be at least 250,000,000 years old, 
a much larger figure than usually given by geologists for 
the period since early Proterozoic. On the basis of these 
figures they may not be more than 300,000,000 years old. 
Pirsson and Schuchert's geology gives 80-90 instead of 
370 million years. 

Many colloidal suspensions, like those in the muddy 
waters of lakes, owe part of their stability to the fact 
that their particles all carry like charges. Of course, the 

G. D. Hubbard— Colloids in Geologic Problems. 103 

movements of the waters help to keep the particles up, 
but in spite of the movements, electrolytes will bring them 
down, so it seems probable that the electric conditions are 
as significant in suspensions as is the agitation of the 
water. If the particles of the colloidal material, and 
other substances in the solution are similarly charged the 
tiny pieces of the colloid do not fall. 

Even when waters are allowed to stand as still as they 
can in a laboratory, some of these colloidal suspensions 
do not clear up in months. It is believed that the parti- 
cles are so small that their own Brownian motion pre- 
vents them from clinging together into flocculent masses 
large enough to fall through the water. Stokes' formula 
gives the following relation: — 

V = 2r 2 g/9k.(d — d') 

V is the velocity of fall of spherical particles of radius r ; 
k is the viscosity of the fluid ; d' and d are the densities 
of the fluid and the particle respectively, and g the accel- 
eration due to gravity. If water and the same kind of 
particles are used on the same table for experiment, the 
variables all drop out but r. 

Theoretically, then, the rate of fall can not quite be 
zero, and permanent suspension becomes impossible if 
the particle is denser than the medium, but practically, 
since V varies with the square of the radius, we can find 
particles so small that their fall would be so slow that 
convection currents and Brownian movement would be 
ample to keep them up. When two or more coagulate 
together, their Brownian kicks are proportionately much 
smaller, and the particle begins to" descend. Protecting 
films of gelatine or other substance which does not mix 
readily with the medium may surround each suspended 
particle and prevent its cohering or coalescing with others 
into masses large enough to fall. Any influence then, 
Brownian motion, similar electric charge, or protecting 
films, may help keep the particles small and aid them in 
staying in suspension. 

Colloids as Cements. — It has long been known that 
nothing soluble occurs in the shales to hold them together, 
and yet they are extremely resistant to pulling strains. 
We say their tensile strength is high. The ultra clay in 
shales and clavs has been shown bv ceramists to consist 

104 G. D. Hubbard— Colloids in Geologic Problems. 

largely of liydrated aluminum silicates, such as kaolin 
and halloysite, also nontronite, iron hydroxides, silicic 
acid, organic matter, and possibly aluminum hydroxide. 
These in their finely divided state are all colloidal, and 
as binders they are stronger when dry than Portland 
cement. They are the cement of shales and the strength 
of some limestones. Colloidal silica also serves as a 
binder in many sandstones, It is known too that the sil- 
ica cements now in the crystalline state were deposited 
in some cases in chalcedonic or colloidal condition. 

In addition to this colloidal material in the shales it 
is known that opaline silica when subjected to the proc- 
esses of metamorphism loses its H 2 and becomes crystal- 
line quartz in a mica schist. Clarke 13 shows that colloidal 
matter contained in most muds and clays is perfectly 
capable of binding under the influence of pressure alone, 
and believes that shales owe most of their coherence to 
unions of this sort. He states on page 609 that it is pos- 
sible for colloidal silica and colloidal hydroxides of alu- 
minum and iron to react and form new silicates when the 
shales are metamorphosed into schists. Thus these finely 
divided materials play their parts in the make-up of 
metamorphic products derived from the shales. 

This process must be closely akin to the one going on 
in the process of binding the materials of our stone roads. 
Grout 14 and Cushman 15 have been carrying on experi- 
ments on the cements of roads, and they think their work 
shows rather conclusively that the binding power of road- 
making materials is due to substances in the colloidal 
state, developed by hydrolyzing the oxides, probably of 
calcium, iron, and aluminum, thus making the body more 
or less plastic. Subsequently by crystallization and 
dehydration they become firm, making a solid road. 

Adsorption and Mineral Colors. — Very much can be 
said on the subject of adsorption and colloids which is 
not geologic, but a number of points certainly belong in 
this paper at this place. Every surface has an attraction 
for other substances. This holding-to a surface is called 
adsorption. Glue on a board or in a beaker adsorbs 
board or glass, that is, there is a strong attraction 

13 Clarke, F. W., Data of Geochemistry, U. S. G. S. Bull. 616, p. 545. 

14 Grout, F. E., Journ. Am. Chem. Soc, 27, 1037, 1905. 

15 Cushman, A. S., U. S. Dep. Agr., Bur. Chem., Bull. 85, p. 92 j Trans. 
Am. Ceramic Soc, 6, 7, 1904. 

G. D. Hubbard— Colloids in Geologic Problems. 105 

between the surface of the one and the particles of the 

Adsorption, then, probably explains some of the colors 
of minerals, inasmuch as the particles of the mineral 
crystal correspond to the surface of the wood or glass, 
and the coloring is in such minute particles as to be in 
the colloidal state. Examples of such intimate relation- 
ships are found in the carbon and iron oxide which give 
the smoky tint to cairngorm, and the carbon which is 
believed to be responsible for the color of amethyst. 
Rose quartz owes its color to the adsorption of colloidal 
titanium on its particles as they come together to build 
up the crystal, and chrysoprase is said to be colored much 
in the same way by nickel oxide. Feldspars have long 
been believed to owe their color to traces of iron oxide 
which seems to be adsorbed by the molecules of the 

The beautiful blue color of some halite has been shown 
to be due to metallic colloidal sodium, and probably the 
bluish tint of sylvite may be traced to the same cause. 
The colors of barite are thought by Patten to be due to 
various oxides. He has shown that it adsorbs salts of 
nickel, cobalt, chromium, iron and manganese. Calcite is 
practically never a colloid, but it shows rather high 
aclsorptive powers for certain colloids. The amber cal- 
cite has been shown to be due to organic matter, and the 
amber fluorite is probably due to hydrocarbons. Ruby 
spinel seems to owe its beautiful tints to adsorbed chro- 
mium oxide, and cerussite in blues and greens to colloidal 
hydrous copper carbonate. More than likely, though this 
is not demonstrated, the soft tints of the zinc salts, cal- 
amine, and smithsonite, from Laurium, Greece, are due 
to the adsorption also of copper carbonates. The colors 
of many gems may probably be due to adsorption of some 
finely divided substance which becomes so intimately 
mixed with the gem materials, yet which occurs in such 
minute quantities, as almost to defy detection. This 
problem has only recently been taken up, and we may, as 
the work goes on, find much more in it than has yet been 

Colloids and the Flotation Process. 1 * "While the flota- 

18 Moses, F. G., Colloids and Flotation, U. S. Bur. Mines, Tech. Paper 200 
furnishes many data for this section. 

106 G. D. Hubbard — Colloids in Geologic Problems. 

tion process now practiced in many of our ore concen- 
tration plants is not really a geologic problem, it is so 
closely connected with geologic materials that I venture 
to call attention to it here. 

In the last few years, many of our large concentration 
plants have installed elaborate facilities for flotation of 
ores. By the use of this process, the heavy sulphides are 
separated from the gangue minerals in what might be 
called a reverse gravity method, because the heavy ores 
come to the top and are from there removed, while the 
lighter materials go to the bottom of the medium. But 
gravity has nothing to do with the process. If galena be 
wetted with water, and a drop of oil be put upon it, the 
oil displaces the water. Galena adsorbs oil much more 
strongly than it does water. On the other hand, if oil be 
spread over a quartz or calcite crystal and a drop of 
water be put upon it, the oil is at once displaced by the 
water. In other words, water is adsorbed much better 
by these common gangue minerals than is the oil. 

For flotation the mineral is finely ground, usually in a 
wet condition, fine enough to pass through a 48-mesh 
sieve. A very small amount of pine, or some other, oil 
with air is beaten into the water which already contains 
the pulp of ore and gangue. The beating must not go far 
enough to make an emulsion or bring the oil into a col- 
loidal state. The oily froth is made up of films of col- 
loidal thinness, but is not emulsified. In this mixture the 
fine particles of sulphides adsorb the oil because it wets 
their surfaces, float to the top by means of their little 
coats, disperse themselves in the froth, and are scraped 
off the flotation tank. Then the froth is beaten out and 
the ore is free from most of the gangue which has, because 
wet by water, gone to the bottom of the flotation tank. 
More than 60,000,000 tons of sulphide ores are thus 
treated in the United States every year and carried much 
more cheaply to a higher concentration than was obtained 
by the old methods. 

In the following ways, then, the subject of colloids 
touches the flotation process. The oil coat held by 
adsorption on the sulphides is so thin that the oil is really 
in colloidal state ; in like manner, films of water of col- 
loidal thinness wet or are adsorbed to the gangues. In 
a few cases colloidal kaolins and other clay substances 

G. D. Hubbard — Colloids in Geologic Problems. 107 

are present, which interfere with, the flotation, by adsorb- 
ing the oil and preventing its nse by the sulphides. Here 
too, the trouble increases if the oil is emulsified, for the 
finer the oil particles are, the more readily do the clay 
colloids adsorb them. Usually the ore is not ground fine 
enough so that its slime is at all colloidal. Electrolytes 
are sometimes used as an aid in flotation, for they help 
to prevent emulsions, or coagulate them if formed. 

Plasticity in Clays. — This subject has been the theme 
for many interesting papers, and of some heated discus- 
sions, but the air seems to be clearing, and Ashley 17 writes 
that it is pretty generally believed by the students of 
ceramics that the control of plasticity in clays is a matter 
of the control of colloids. A number of other writers 
can also be quoted in support of this position. A French 
ceramic chemist, T. Schlosing, in 1888, P. Rohland in 
various papers from 1902-9, A. S. Cushman, quoted above, 
and F. W. Clarke in the Data of Geochemistry, all agree 
that the plasticity of clays can be very directly traced to 
their colloid content. Schlosing shows, according to Ash- 
ley, that the amount of colloidal material in the best clays 
is small, and rarely exceeds one and one-half per cent. 
As the percentage of material in the colloidal state 
decreases, so does the plasticity, and a clay with one-third 
of the amount mentioned is a lean clay. Likewise, if the 
amount is greatly increased above the one and one-half 
per-cent, the clay becomes less plastic, and more sticky. 

It is well known that the plasticity of clays is destroyed 
by ignition and that the colloids are also made to change 
state by the same treatment. On the other hand, many 
thoroughly ignited clays absorb water quite as well as 
unignited ones. This makes it clear that the colloidal 
material is not the cause of the absorption of water, nor 
is the absorption of water vitally related to plasticity. 
Series of experiments have been made with lean clays to 
increase their plasticity, and it has been found that very 
poor clays can be made plastic enough to be worked suc- 
cessfully by adding small quantities of certain colloids; 
agar-agar .08% increases the plasticity of different clays 
as much as 40 and in some cases 60%, and aluminum 
cream produced the same effects by adding in much larger 
quantities, e.g. about 3%. 

17 Ashley, H. E., Bur. Stand., Tech. Paper 23, 1911. 

108 G. D. Hubbard— Colloids in Geologic Problems. 

As was shown on a previous page, colloidal material 
increases the tensile strength of shales, and it has been 
shown to do the same for clays. Other colloidal ^ sub- 
stances have also been used to increase the plasticity. 
Humus, for example, an organic colloid, produces bene- 
ficial results in lean clays. Clays in a cool, moist place 
improve in plasticity even within a few weeks. It is 
believed that the development of organic matter in the 
clay by the growth of bacteria, or even protozoa, is 
responsible for the improvement, for the protoplasm of 
these minute organisms is itself a colloid. 

If plasticity of clays should be found to be dependent 
upon the inorganic colloids in them, it might be pertinent 
to ask if clays and shales laid in the sea would be more 
plastic than those laid in fresh water, where the finest or 
colloidal material had difficulty in being precipitated. 

Ries is has shown that the fineness of the material, the 
thinness of the plates in the shale, and the colloids, are 
each in themselves inefficient to fully explain plasticity. 
His theory suggests that plasticity may be due to cohesion 
and adhesion factors which depend on the constitution of 
the molecule, but not on the chemical composition. It 
might be said that he does not seem to be fully satisfied 
with his explanation, for he adds that "practical work 
to improve the plasticity of clays may well follow lines 
already started, such as the addition of colloids, and 
weathering, which may mean the addition of colloidal 
material by bacterial growth. ' ' 

One can never understand the raw materials and their 
relations to the finished products in cements, brick and 
terra-cotta, pottery, porcelain, enamels, and glass, until 
he has dipped rather seriously into the study of colloids 
from the geologic side. Weathering is a colloid-pro- 
ducing process, as well as a maker of most of the soils and 
the salts of the sea. Since macadam, brick, cement, and 
asphalt roads all start with materials whose vital proper- 
ties are connected with the colloidal state of matter and 
the geologic processes that produce these materials, the 
road industry has real need of a colloid geologist. 

Colloids in Soils. Perhaps this topic will take us as 
far into questions of physiography as the last has into 
metallurgical and industrial processes, but in the last 

ls Kies, H.,.Geol. Survey, W. Va., vol. 3, pp. 46-54, 1905. 

G. D. Hubbard— Colloids in Geologic Problems. 109 

analysis, all three go straight back to geologic processes, 
and involve geologic problems. Soils result from the 
normal geologic decay of rocks, primarily of silicate 
rocks. In their decay, the elements K. Xa. Ca, and Mg 
usually go into true solution in their secondary salts, 
while silica, aluminum, and iron go chiefly into colloidal 
solution and constitute the ultra-clay material of clays 
and shales. 

In solutions, the colloidal material can be separated 
from the crystalline material by the use of the Sharpie's 
centrifugal machine, capable of producing a force seven- 
teen thousand times that of gravity, or by the use of the 
Pasteur-Chamberland filter. This separation can also be 
brought about by coagulation by the addition of salts, 
and by dialysis. 

In soils the body is essentially sand and clay, the sand 
being made of fragments of many kinds of minerals, but 
mostly of quartz : and the clays mostly of hydrous 
aluminum silicates, with smaller quantities of aluminum 
and iron hydroxides. "When the salts mentioned above 
come to the soils they are carried on through, pro- 
viding the water has free circulation and drainage 
below : but if there is insufficient rainfall to equal evapo- 
ration, then these salts may be left in the soils and be 
continually carried to the surface by evaporation of the 
water. The colloidal material, however, will usually be 
adsorbed and will remain in the soil. Too much of the 
latter tends to clog a soil and prevent the free and neces- 
sary movement of air and water. 

This last item becomes particularly troublesome in soils 
that must be irrigated, for irrigation waters differ from 
rain waters in carrying both true solution and colloidal 
solution materials, thus furnishing more material to clog 
the soil than rain waters. The difficulty is still further 
increased by the fact that most of the water on irrigated 
lands is removed by evaporation so that everythii 
both kinds of solution is left in the soil. The colloids 
become a nuisance usually much before the salts do when 
ordinary stream water is used for irrigation. The col- 
loids tend to cement the soil together some little distance 
below the surface, usually not beyond the reach of the 
plow, and produce there a "hard pan" layer. This inter- 
feres with the -movement of the water, either up or down. 

110 G. D. Hubbard — Colloids in Geologic Problems. 

and of course prevents the roots striking deeply. It can 
be broken np by deep cultivation, but the soil experts 
are now of the opinion that treatment with an electrolyte 
is really better. Aluminum sulphate has been success- 
fully used in a number of cases. It acts much as the salts 
do in the sea, by coagulation of the colloidal material into 
little pellets, large enough so that the water and the air 
can get among them, and thus prevents their operation 
as cements. 

In western United States, where most of our irrigation 
is carried on, there have been discovered in recent years 
great quantities of aluminum sulphate, and it is believed 
that a large use of this salt will greatly extend the life 
of our irrigated soils. 

The subject of colloids in geology is just beginning to 
attract the attention of men who should be concerned with 
it. While the chemists have tackled their colloid prob- 
lems with vigor and enthusiasm the geologists have all 
but neglected the whole field. There should be many 
geologists turning their attention seriously to the solution 
of the great numbers of problems now before us, such as 
these suggested in this paper, and there should be many 
more geologists keeping up with the literature that dis- 
cusses these colloid problems. 

Department of Geology and Geography, 
Oberlin College, Oberlin, Ohio. 

R. S. Lull — Primitive Pecora in Yale Museum. Ill 

Art. IX. — Primitive Pecora in the Yale Museum; by 
Richard S. Lull. 

[Contributions from the Othniel Charles Marsh Publication Fund, Peabody 
Museum, Yale University, New Haven, Conn.] 

The genus Hypertragulus was established in 1873 by 
Professor Cope to include a group of deer-like creatures 
which in some respects resembled those of the genus 
Lept ornery x very closely. The main distinctions as 
shown by the dentition are as follows : 

Hypertragulus possesses : the superior laniary canine, 
P 1 , separated from both C 1 and P 2 by diastemata, upper 
molars without mesostyle, and M 3 bearing three ribs on 
its outer face, all of which are the converse of Lepto- 
meryx. The lower dentition differs from that of Lepto- 
meryx in the development of a caniniform P x , a diastema 
behind P 2 , a compressed and elevated P 3 which is shorter 
than the three-lobed P 4 , and in the fact that in M 3 the 
posterior crescents are opposite each other, subequal in 
size, and not separated posteriorly by a fissure. 

The geological range of Hypertragulus is apparently 
from Middle Oligocene Oreoclon beds to Lower Miocene 
Rosebud. Geographically, the genus is found in the 
Great Plains region — Colorado to South Dakota, Mon- 
tana, Canada, and the John Day Basin of Oregon. 

The following species have been named, H. calcaratus 
Cope from the Great Plains region being the type of the 

H. calcaratus Cope 1873. 

H. tricostatus Cope 1873. 

H. transversus Cope 1889. 

H. hesperius Hay 1902. 

H. ordinatus Matthew 1907. 

Allomeryx planiceps Sinclair 1905. 

To which are added in this paper : 

H. minutus, sp. nov. 
Leptomeryx obliquidens, sp. nov. 
Xanotragulus loomisi, gen. et sp. nov. 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 20.— August, 1922. 

112 R. 8. Lull — Primitive Pecora in Yale Museum. 

Hyperiragulus calcaratus Cope. 

The original description of the genns is brief, and two 
species, H. calcaratus and H. tricostatus, are mentioned 
without definition, the first adequate description appear- 
ing in Paleontological Bulletin No. 16, p. 7, under the 
genus Leptauchenia. In this description Cope mentions 
several characters which are really of generic value, the 
real specific distinctions of H. calcaratus being apparently 
as follows : 

Molars bear a slight external cingulum which has a 
small cusp between the two outer crescents, representing 
either an incipient or vestigial mesostyle. M 3 relatively 
somewhat smaller in TI. calcaratus, and the metastyle, 
while variable, never as pronounced as in the John Day 
form. In general, the teeth are somewhat smaller than 
in the John Day jaws, and in one instance where P 3 is 
preserved, it is markedly smaller. Interrupted cingula 
are present in all the Hypertragulus jaws before me, 
although omitted from every published figure. 

Cope's measurements are: 


Length of five molars 26 

Length of three true molars 17.5 

Length of last true molar 8 

Width of last true molar 7 

This species is smaller than the smallest of the genus 
yet described (Cope). 

Of the Yale material, assuming Cope's figures to be 
correct, but one specimen agrees exactly with the type 
in the combined length of the ^ve teeth, and even in this 
specimen M 3 is only seven-eighths as long. The third 
molar, however, has been shown to be variable and the 
degree of wear also affects its dimensions. I find no 
evidence in the Yale Great Plains material of even sub- 
specific departure from H. calcaratus. 

Hypertragulus tricostatus Cope. 

Cope thought he recognized a second species about the 
size of H. calcaratus but distinguished therefrom by the 
presence of but three ribs on the outer side of the third 

R. S. Lull — Primitive Pecora in Yale Museum. 113 

molar, the characteristic heel being absent. This molar 
also lacked the posterior cingulum. H. tricostatus is 
considered a synonym of H. calcaratus by Hay, and with 
this the present writer agrees, for ont of twenty-four 
individuals in the Yale Collection represented by the 
upper molar teeth, the metastyle of M 3 varies from good 
development to marked reduction in at least four speci- 
mens, but is never entirely absent. The degree of devel- 
opment of the posterior cingulum is also variable to total 
obsolescence. As tricostatus is founded upon a single 
specimen, it seems to represent merely the extreme of a 
variational series of which the means were still extant 
and therefore not a separate species but a varietal ten- 
dency. Cope himself says in 1884 (p. 24) : "I know but 
the one species, the H. calcaratus Cope," a statement 
which, as Matthew rightly says, invalidates J?, tricostatus. 

Material from the John Day Formation. 

Hypertragulus hesperius Hay. 

Both Cope and Leidy have discussed hypertragulids 
from the John Day Basin of Oregon, but have referred 
them either to H. calcaratus or to Leptomeryx evansi 
without further attempt at specific differentiation. Scott 
has also figured John Day material belonging to this 
genus under the name H. calcaratus. Hay, however, in 
his catalogue, p. 675, gives the new name H. hesperius to 
the John Day hypertragulids, with neither definition, indi- 
cation of type, nor restriction to any one level. Doctor 
Matthew informs me (personal communication) that Cat. 
No. 7918, A. M. N. H., the specimen figured by Cope and 
later by Scott, is to be regarded as the type. It consists 
of the skull and jaws and a few fragments of the skeleton. 
The matrix is "greenish tinged with buff and rather 
hard" but there is no record of its exact level or locality, 
save that it is from the John Day. The matrix color, 
judging from the known distribution of the large amount 
of John Day material in the Yale Collection, would indi- 
cate middle John Day as the horizon of this type. 

Cope's description of H. hesperius gives no specific 
characters other than that the size is the same as H. cal- 
caratus, sometimes distinctly larger; later (1884, p. 25) 

114 R. S. Lull — Primitive Pecora in Yale Museum. 

he says "I can not distinguish the John Day species from 
the H. calcaratus, although the size is generally distinctly 
larger." He does, however, use the John Day material 
for his description of the morphology of the feet of the 
genus, but this part of his description is generic and not 

The type skull is fairly complete, except that it is 
injured in the pterygoid region, and represents a fully 
mature animal in which the tooth pattern of M 1 is entirely 
obliterated. Anteriorly, a portion of the large canine 
alveolus is present, but no trace of the premaxillaries is 
preserved. There is no jugal process, but the postorbital 
process of the frontal forms about half the posterior 
margin of the orbit. Scott's figure, which was drawn by 
Mr. R. Weber, is the more accurate of the two except for 
the restoration of the superior canine behind the canini- 
form P 1? which the author evidently mistook for the true 
lower canine tooth. Scott further restores the orbit as 
though it were closed behind, for which there is no 

Allomeryx planiceps Sinclair. 

This form agrees with Hypertragulus in the absence of 
the mesostyle on the molars and in the development of 
the metastyle on M 3 . It is distinguished by the closure 
of the orbit behind by the frontal and jugal processes 
which overlap but are not completely fused. The bullae 
are small and separated from the basi-occipital by an 
outgrowth of the petrosal. The brain-case is shorter 
than in Hypertragulus, and the interorbital tract and 
sagittal crest lie in one plane. 

Merriam and Sinclair in a later paper (1907) doubt the 
generic rank of these several characters. The closure of 
the orbit, however, is an evolutionary advance which is 
significant, especially as in the known species of Hyper- 
tragulus the jugal process is undeveloped. 

An imperfect skull, No. 10227, Y. P. M., shows Allo- 
meryx characters in so far as preserved, especially those 
noted in the basicranial region. It also possesses an 
ample brain-case. Unfortunately the postorbital region 
is not preserved. The Yale skull is from the upper John 
Day beds. 

R. S. Lull — Primitive Pecora in Yale Museum. 115 

Hypertragulus ordinatus Matthew. 

This species is based on a complete lower jaw (No. 
13011, A. M. N. H.) from the lower Rosebud beds on 
Porcupine Creek, South Dakota. It is larger than the 
type species of the genus, H. calcaratus, and about equal 
to H. liesperius. It may be distinguished from either by 
the following characters : Closing of diastema between 
P 2 and P 3 , and great reduction of diastema between C± 
and Pj ; increased length of diastema between caninif orm 
P x and P 2 , so that general proportions of jaw are about 
the same ; molar crowns more hypsodont than in H. lies- 
perius, about as in H. calcaratus ; P x and P 2 shorter and 
proportionately higher than in the John Day species but 
less reduced than in that from the White River. In the 
upper Rosebud this species is replaced by Blastomeryx. 

Hypertragulus minutus, sp. nov. 

Holotype, Cat. No. 10545, Y. P. M. Upper Oligocene (upper John Day), 
Oregon. Fragments of upper and lower jaws and teeth. 

A very small form, apparently Hypertragulus. Dis- 
tinguishable from H. liesperius mainly by its small size. 
Cingula well developed, but metastyle of M 3 much 
reduced, not forming the conspicuous "heel" of lies- 
perius. Its measurements are, compared with those of 
liesperius : 

H. minutus H. hesperius Yale spm. 
mm. mm. 

Length, M 1 to M 3 14 20 

Length, M 3 5.8 8.5 

Width, M 3 5 8.5 

Leptomeryx obliquidens, sp. nov. 

Holotype, Cat. No. 10541, Y. P. M. Oligocene (Protoceras beds), Her- 
mosa, South Dakota. Poorly preserved skull and jaws, with entire series 
of cheek teeth. 

Distinguishing characters. — Large size; superior 
molars obliquely set ; temporal ridges meeting at a wide 
angle to form the sagittal crest which is sharp and thin 
throughout its preserved length; bullae laterally com- 
pressed, elongated oval in shape. 

116 R. 8. Lull — Primitive Pecora in Yale Museum. 

This form is distinguished from L. evansi of the Oreo- 
don beds, aside from its geological level, by being about 
one fourth larger, and by the great obliquity of the 
molars. The bullae in all specimens of L. evansi before 
me are more inflated. The type is an aged animal, hence 
in their present state of wear there is little that is dis- 
tinctive about the teeth. The general shape of the man- 
dible and the position of the mental foramina, of which 
there are two, are about the same in both species. 
Cingula are absent and the external basal pillars are 
feebly developed, as in evansi. 

Measurements are as follows: 

L. evansi 
L. obliquidens Cat. No. 10542 

Holotype Eatio Y. P. M. 
mm. mm. 

Length, tooth-row, P 2 to M 3 (left) . . 48 1.23 39 

Length, M 1 to M 3 26 1.24 21 

Length, M 3 9 1.20 7.5 

Width, M 3 12 1.54 7.7 

Width, P 2 5 1.25 4 

Length, tooth-row, P, to M 3 51.7 1.21 42.5 

Length, M ± to M 3 31 1.26 23.7 

Length, M 3 14 1.29 10.8 

Width, M 3 - 7 1.40 5 

Nanotragulus loomisi, gen. et sp. nov. 

(Fig. 1.) 

Holotype, Cat. No. 10330, Y. P. M. Collected by E. S. Lull in 1908 at 
Castle Butte, Big Muddy Eiver, near Spanish Mines, Wyo. Miocene (lower 
Harrison). Palate with upper cheek teeth complete, right lower jaw fairly 
complete, with six teeth, left with series of four teeth. Detached premolar. 
Imperfect petrosal? 

Distinguishing characters: Upper dentition. — Teeth 
subhypsodont ; P 2 -M 3 a compact series withont diaste- 
mata; mesostyle lacking on molars; no trace of molar 
cingula nor of internal basal pillars ; molars simple, of 
f our crescents each ; M 3 with four external pillars ; meta- 
style enlarged toward base of tooth. Premolars three 
only, preserved in situ ; P 2 trenchant, single-cusped, two- 
rooted, slightly grooved on outer face. P 3 also trenchant, 
of greater antero-posterior diameter than P 2 , apparently 
three-rooted, with postero-internal cingulum but no 
internal cusp (deuterocone), and with two faint external 

R. S. Lull — Primitive Pecora in Yale Museum. 117 

ridges. P 4 a unique tooth, triangular in section, with 
outer crescent not fully evolved, flanked by three faint 
ridges on its outer face ; postero-internal cingulum form- 
ing a sharp, straight ridge abutting against the well devel- 
oped deuterocone but distinct therefrom in the unworn 
tooth. Measurements : 


Length P 2 to M 3 27 

Length, M 1 to M 3 17 

Length, M s 7 

Width, M 3 , at base of ant. crescent 5 

Lower dentition. — P 3 -M 3 a continuous series. Prob- 
ably slight diastema separating P 2 and P 3 , but condition 
of specimen renders this uncertain. Molars simple, with 

Fig. 1 


Fig. 1. — Nanotragulus loomisi, gen. et sp. now Holotype. A, left upper 
dentition. B, right lower dentition. X a little more than 3. 

slightly developed cingula on anterior face. No acces- 
sory pillars. Posterior column of M 3 as in Hypertrag- 
ulus, with outer and inner lobes opposite and not 
separated posteriorly by a cleft as in Leptomeryx. P 2 a 
simple, compressed cone, two-rooted, slightly recurved. 
P 3 high-crowned, but not so much so as in Hypertragulus, 
hardly rising above P 4 , with a main anterior cusp and 
rather prominent heel. P 4 the most complex, laterally 
compressed, high protoconid, and pronounced heel. 
Antero-internal cusp well developed. Protoconid flanked 
on inner side by a second cone which is confluent with it, 

118 B. S. Lull — Primitive Pecora in Yale Museum. 

while a third bears the same relation to the heel. Thus 
there are three internal buttresses separated by valleys. 
No traces of cingula on premolars. Measurements : 


Length, P to M 3 ca. 28* 

Length, M^ to M 3 18.3 

Length, M 3 8 

Width, M 3 3.5 

*Slightly elongated by fracture between P, and P 3 . 

A detached small, slender, two-rooted premolar is 
present. It probably represents P 2 of the left ramus and 
is not distinguishable from that already described. 

Relationships. — This is the smallest artiodactyl thus 
far described from the Lower Miocene. In this respect 
it is suggestive of the Oligocene Hypisodus, with which 
it also agrees in hypsodonty and in the absence of the 
mesostyle. They differ, however, in the absence of a 
buttress on the posterior external crescent of M 1 - 2 and in 
the character of the premolar teeth. The inner anterior 
crescent of M 1 is markedly different. 

The form under discussion differs from Leptomeryx 
in the absence of mesostyles, character of heel of poste- 
rior lower molar and of the premolars, and in size. "With 
Hypertragulus it agrees in the absence of mesostyle and 
the character of the heel of M 3 , but differs again in its 
much smaller size and the character of its premolars. 
Nanotragulus differs markedly from Merycodus in size 
and geologic level, also morphologically in the absence in 
the former of the mesostyle, in the convergence of the 
molar crowns, and in the character of the premolar teeth. 
They agree mainly in hypsodonty. 

From its contemporary, Stenomylus, the new genus 
also differs very markedly in size and in the character of 
the premolars. The two genera agree, however, in 
hypsodonty, and in the absence of mesostyle, though the 
latter is indicated on M 1 of Stenomylus. In Stenomylus, 
moreover, the external face is relatively smooth except 
for parastyle and metastyle. In the lower dentition of 
Stenomylus there is greater simplicity, especially in P 4 , 
which in no way resembles that of Nanotragulus. In 
other words, Nanotragulus is deer-like, not camel-like as 
is Stenomylus. 

R. 8. Lull — Primitive Pecora in Yale Museum. 119 

I can not at present place the new genus elsewhere than 
in the Hypertragulidse, but it is not clearly derivable from 
any known Oligocene form except possibly Hypisodus. 

The generic name Nanotragulus refers to its dwarfed 
size, while the species is named for Professor F. B. 
Loomis of Amherst College, leader of the joint Amherst- 
Yale expedition of 1908 during which the type was col- 
lected, and the first to recognize its unique character. 


Cope, E. D. 1873A. [On Menotherium lemurinum, etc.] Proc. Acad. Nat. 

Sci., Philadelphia, 25, 419-420. 
— 1873B. Third notice of extinct Vertebrata from the Tertiary of the 

Plains. Pal. Bull. 16. 
— 1884. On the structure of the feet in the extinct Artiodactyla of North 

America. Proc. Amer. Philos. Soc, 22, 21-27. 
—1889. The Artiodactyla. Amer. Nat., 23, 111-136. 
Leidy, Joseph. . 1873. Contributions to the extinct vertebrate fauna of the 

Western Territories. Kept. U. S. Geol. Survey Terr., 1, 14-358. 
Matthew, W. D. 1907. A Lower Miocene fauna from South Dakota. Bull. 

Amer. Mus. Nat. Hist., vol. 23, 169-219. 
Merriam, J. C. and Sinclair, W. J. 1907. Tertiary faunas of the John 

Day region. Univ. Calif., Bull. Dept. Geology, vol. 5, 171-205. 
Scott, W. B. 1899. The selenodont artiodactyls of the Uinta Eocene. 

Trans. Wagner Free Inst.. Sci., Philadelphia, 6, 1-121. 
Sinclair, W. J. 1905. New or imperfectly known rodents and ungulates 

from the John Day series. Univ. Calif., Bull. Dept. Geology, vol. 4, 


120 Diener — Critical Phase in History of Ammonites. 

Art. X. — A Critical Phase in the History of Ammonites ; 
by C. Diener. 

The extinction of ammonites, those masters of the 
Mesozoic seas, near the close of the Cretaceous period 
is a fact well known to all students of palaeontology. 
The number of their families and genera is diminishing 
gradually during the Senonian epoch. Five species only 
reach into the stage of the Maestrichtian. Not one passes 
the fatal border of the Danian. 

It is, however, less known, that the existence of ammon- 
ites was threatened by a similar crisis at a considerably 
earlier period of the Mesozoic era. They passed through 
a very critical phase at the boundary of the Ehaetic and 
Liassic stages. All but one phylum of Triassic ammon- 
ites became extinct at the close of the Ehae tic epoch. By 
the survival of this single phylum, which in the Lower 
Lias gave rise to the development of a new and rich fauna, 
the ammonites were saved from complete extermination. 

E. v. Mojsisovics was the first to notice this remarkable 
crisis in the history of Triassic ammonites. For fuller 
details the reader is referred to J. F. Pompeckj, "Am- 
moniten des Ehaet" (Neues Jahrb. f. Mineral., etc., 
1895/11, pp. 1-46) and to some of my own memoirs. 

The Upper Triassic deposits of Tethys are divided 
generally into three subdivisions, the Carnic, Noric and 
Ehaetic stages. E. v. Mojsisovics divided both the Carnic 
and Noric stages into three substages, thus imparting to 
the Ehaetic stage a taxonomic value inferior to that of 
the two preceding ones. Many genera belonging to all the 
known families of Upper Triassic ammonites reach 
the acme of their development in the Carnic stage. 
Although a considerable number of older genera are 
found for the last time at this level, the ammonite fauna 
of the Noric stage is a continuation and evolution of the 
Carnic fauna in every branch of life. The last life phase 
of the Noric stage seems to be the first which is distin- 
guished from the preceding by the apparent extinction 
of numerous wide-spread and important genera and by 
the absence of any new elements either of foreign origin 
or derived from endemic forms. Nevertheless it is doubt- 
ful whether a single family of lower Noric ammonites 
becomes really extinct. 

Diener — Critical Phase in History of Ammonites. 121 

This decay is completed in the Noric epoch. From this 
stage eleven forms of ammonites only have been enumer- 
ated by Pompeckj, all of them of decidedly Triassic 
affinities. Five belong to the Noric genus Choristoceras 
and its subgenus Peripleurites, a phylogerontic descen- 
dant of the Ceratitidae, whose last whorl became gradually 
uncoiled. Arcestes, the true leading genus of the Hall- 
statt limestone, is still represented by two species. To 
these are added one species of Monophyllites (Mojsvar- 
■ites), of Megaphyllites, and a specifically undeterminable 
representative of Cladiscites, all genera of considerable 
vertical range. A single newcomer is indicated by Hes- 
perites, a genus still imperfectly known, which is probably 
allied to the family of Trachyceratidae. 

It is noteworthy that not a single ancestral represen- 
tative of Liassic ammonites is recognized in this assem- 
blage. The discovery of one other genus is to be expected 
beds of Rhaetic age. This is Phylloceras or, more exactly, 
with certainty, although it has as yet not been found in 
Rliaeo phyllites , if this subgeneric designation is extended 
to all widely umbilicated species of Phylloceras. Rhaco- 
phylUtes debilis Hau. and R. neojurensis Quenst. are 
among the most common leading fossils of the upper 
Noric substage. In the Lower Lias, Phylloceras, Rhaco- 
phyllites and Euphyllites are remarkable for their rich- 
ness and variety. The apparent intermittence of Phyl- 
loceras- in the Rhaetic is tkeref ore purely accidental. It 
is in reality tke only genus surviving tke general exter- 
mination of Triassic ammonites. 

Tke importance of tke gradual decline of Triassic 
ammonites during tke Rhaetic epoch is evident from a 
comparison with the number of genera in the Carnic and 
Noric faunae. Those faunae do not contain less than 146 
genera and subgenera of ammonites, which were reduced 
to six in the Rkaetic stage. Hyatt was certainly rigkt in 
speaking of a "culmination of ammonites in tke Upper 
Trias after a period of uninterrupted progressive evolu- 
tion from tke early Devonian.'' Botk tke Carnic and 
Xoric ammonites were highly varied, including forms 
witk long and short body-chambers, with few and simple 
clydonitic sutures (Lobiies) and with a very large number 
of the most complicated sutural elements (Pinacoceras) ; 
smooth, globose shells with serial lobes (Arcestes) and 

122 Diener — Critical Phase in History of Ammonites. 

extremely flattened shapes (PompecJcjites) ; shells 
exhausting* almost every possible combination of sculp- 
ture from the most graceful ornamentation (Acanthin- 
ites) to stout ribs (Heraclites) and profusely tuberculated 
costations (Trachyceras). 

The close of the Rhaetic epoch is marked by the final 
disappearance of all Triassic types, excepting Phyl- 
loceras. Primitive and highly specialized forms were 
equally subjected to this general extermination. 

In the eastern Alps the beds of the lowest Lias follow 
above those containing a Rhaetic fauna without any uncon- 
formity. There is no trace of a hiatus nor of any 
diastrophic movement between the two groups. Never- 
theless the ammonite fauna of the lowest zone of the 
Mediterranean Lias is entirely different from that of the 
Upper Trias. The first impression of this Liassic fauna 
is the sudden introduction of a large number of types 
which are only a little less manifold and diversified than 
those of the Upper Noric, but do not exhibit any phylo- 
genetic affinities with them. We are indebted to F. 
Waehner for their careful and detailed examination. 

There is little doubt that the extinction of the different 
phyla of Triassic ammonites prepared the way for the 
evolution of a new and vigorous stock, which originated 
from the genus Phylloceras, the only one which connects 
the faunae of the Triassic and Liassic periods. Phyl- 
loceras is the ancestor of the two leading families of the 
lowest Lias, the Arietitidae and Lytoceratidae. Waehner 
and Pompeckj have demonstrated their intimate relation- 
ship with Psiloceras, the most primitive element of the 
Arietitidae. Together with Psiloceras, more specialized 
types of the Arietitidae : Mgoceras, Schlotheimia, Ariet- 
ites, make their appearance in the deepest zone of the 
Lias. But they are comparatively rare, Psiloceras 
remaining the predominant genus in this and the follow- 
ing life-phase. All these genera are linked together most 
closely with the ancestral Psiloceras. 

Of equal moment is the sudden appearance of the Lyto- 
ceratidae in the Lower Lias, where they are represented by 
the genera Lytoceras, Ectocentrites and Pleuracanthites. 
Forms transitional between Pleuracanthites and Psilo- 
ceras have been described by Waehner. Thus Phyl- 

Diener— Critical Phase in History of Ammonites. 123 

loceras was destined to give rise to all Lower Triassic 
ammonites by the intervention of Psiloceras. 1 

Tims an aspect quite different from that of the Upper 
Trias is given to the ammonite fauna of the Lower Lias. 
Not one of the numerous and diversified genera of world- 
wide distribution, belonging to the families of Arcestidse, 
Cladiscitidse, Pinacoceratida?, Haloritidae, Tropitidae, 
Didymitidse, Ceratitidae, Tirolitidse, and Trachyceratidse is 
represented in the latter. Their place has been taken by 
Arietitida? and Lytoceratidae. Phylloceras, which never 
played an important part in the fauna of the Upper Trias, 
was the only survival and was destined to become the 
ancestor of all Liassic ammonites. 

In direct opposition to these facts, Steinmann denied 
the extermination of Triassic ammonites at the close of 
the Ehaetic epoch. His reconstruction of a phyletic tree, 
in which Macrocephalites is branching off from Juvavites, 
Sphceroceras from Halorites, Harpoceras from Disco- 
tropites, Desmoceras from Arcestes, P a cliy discus from 
Cladiscites, need not be discussed here. It means toying 
with possibilities, the reality of which can never . be 
proved. 2 One of his critical arguments, however, 
deserves consideration. He believes the paheontological 
record not to be sufficiently perfect to prove a real decline 
of the Triassic ammonites during the Silastic epoch. It 
is true that cephalopod-bearing strata of Ehaatic age have 
scarcely been discovered up to now outside the north- 
eastern Alps. But here they are as rich in ammonites as 

1 In connecting Psiloceras with Phylloceras (Bhacophyllites) I am fol- 
lowing J. F. Pompeckj ? s view, which has been set forth by this author in his 
memoirs, "Xote sur les Oxynoticeras clu Sinemurien du Portugal, etc." 
(Comm. serv. geol. Portugal, VI, 1906-1907, p. 332) and "Zur Kassenper- 
sistenz der Ammoniten' ; (3. Jahresber. d. niedersachs. Geol. Ver. Hanno- 
ver, 1910, p. 82). E. v. Mojsisovics prefers to consider a specialised type 
of MonophylJites (Mojsvarites phuwrboides Winkler) as the ancestor of 
Psiloceras. Winkler's description and illustration, on which this suggestion 
has been based, are not absolutely reliable, and the type-specimen itself has, 
unfortunately, been lost. 

It makes, however, little difference, whether the one or the other view is 
adopted, Mojsvarites itself being closely related to the Phylloceratida?. 
According to Pompeckj, one genus only, Phylloceras, persists throughout the 
Triassic and Jurassic periods. In following E. v. Mojsisovics we have to 
record, simultaneously with the decline of Monophyllites, the first appearance 
of a new and transitional form, connecting this genus with Psiloceras, the 
undoubted ancestor of all Arietitida. 

-Its onlv advocate is O. Wilckens (Xaturwiss. Wochenschrift, X. F., X., 
Xo. 45, Jena, 1911, p. 20). 

124 Diener — Critical Phase in History of Ammonites. 

many beds of the Ladinic or Noric stages. Our knowl- 
edge of Rhaetic ammonites is certainly not more limited 
than that of Permian ammonites after the discovery of 
the Artinsk and Sosio fauna. There is, consequently, as 
much evidence of a decline of the group during the Rhsetic, 
as there is of a decline of the trilobites during the Car- 
boniferous and Permian. 

Such are the facts. They show us a great dying-out 
of ammonites towards the close of the Triassic and a 
rebirth, as it were, of a new fauna in the early Liassic, 
giving rise to the great wealth of Jurassic ammonite 
evolution. In entering into a discussion of the probable 
causes of this remarkable event in the life history of 
ammonites, we have to face the grave problem of the 
repeated extinction of large and flourishing groups of 
organisms. That this extinction has been partial only, 
affecting all but one stock of Triassic ammonites, marks 
the special case of our problem. 

If we reflect on the multitude, the variety, and the 
complexity of the facts to be explained, and the scantiness 
of our information regarding them, we shall be ready to 
acknowledge that a full and satisfactory solution of so 
profound a problem is hardly to be hoped for, and that 
the most we can do in the present state of our knowledge 
is to hazard a more or less plausible conjecture. 

In discussing the possible causes of the decay and final 
extermination of Upper Triassic ammonites, it will be 
best to follow the lines which have been traced by H. P. 
Osborn in his memoir on the causes of extinction of 
Mammalia (Amer. Naturalist, XL, 1906, p. 769). 

Changes of Geographical Conditions. — The Triassic 
was on the whole a geocratic (land) period. A transgres- 
sion of marine Rhaetic beds is confined to the coasts of 
western Europe. It is counterbalanced by a regression 
of the sea in southern China, Japan, North and South 
America, where the Rhaetic stage is represented by 
deposits of terrestrial and lacustrine origin only. There 
is but a small change in the distribution of land and sea 
during the Lower Lias, which in some regions of the 
ancient geosynclines is marked by a transgression of a 
rather limited range. The extinction of Triassic ammon- 
ites consequently does not admit of an explanation by 
changes of geographical conditions. 

Changes of Climate.— The importance of this factor has 

Diener — Critical Phase in History of Ammonites. 125 

been advocated very strongly by C. Schuchert, 3 who 
insists on a general lowering of the temperature during 
the Liassic period, chiefly on the strength of the argu- 
ments of Handlirsch. I am happy to agree with this 
learned author in the opinion that the facts which prove 
the influence of climatic changes are many and weighty, 
but I think that the extinction of Triassic ammonites at 
the close of the Rhaetic does not admit of this explanation. 

It can hardly be too often repeated that the decay of 
ammonites, near the close of both the Triassic and Creta- 
ceous periods, was gradual, that in the first instance it 
clearly began in the Upper Noric and continued through- 
out the Rhaetic. Now, we are well informed about the 
climatic conditions of Upper Noric time, due to the dis- 
covery of rich faunae of reef -building corals in this sub- 
stage. Such faunas are known to us from the Austrian 
Alps, from Timor, Nevada, Oregon and Alaska. J. Per- 
rin Smith demonstrated the identity of nearly all his 
Alaskan species with types from the Upper Noric Zlam- 
bach beds of the Salzkammergut. The presence of this 
fauna under the 60th degree of north latitude is contra- 
dictory to the suggestion of a lowering of temperature in 
the Upper Noric seas. It may be equally well to call 
attention to the wide distribution of several species of 
ammonites and bivalves (Pseudomonotis ochotica) 
throughout the Pacific Ocean and into the arctic region 
(New Siberia), which is in favor of a comparatively 
equable but not of a low temperature. 

Nor does the flora of the Rhaetic stage exhibit any traces 
of increasing cold. It is of a remarkable uniformity in 
North America, England, Sweden, Germany, eastern 
Greenland, Spitzbergen, Persia, India, Japan, China, New 
South Wales, New Zealand, South Africa, Argentina, 
Chile and Honduras, and seems to prove a climate more 
uniform and milder in the polar regions than that of the 
present day. 

If a period of cooling set in at the close of the Rhaetic 
epoch — and I do not dissent from Professor Schuchert 's 
opinion in this respect — it came too late to influence the 
extermination of Triassic ammonites, for this had been 
heralded long before by their gradual decay. 

3 Ch. Schuchert, Climates of geologic time, Carnegie Inst. Washington, 
Publ. Xo. 192, p. 284. 

126 Diener — Critical Phase in History of Ammonites. 

Lack of Internal Adaptation and Inadaptability. — 
Arcestes is one of the most persistent ammonite genera, 
ranging from the Anisic into the Khsetic stage, without 
undergoing any modifications of its characters. Thus its 
inadaptability can scarcely have been the primary cause 
of its extinction. 

In the family of Ceratitidae Hyatt signalized the first 
symptoms of general regression by the appearance of 
uncoiled and turriliticonic genera in the Noric stage. 
The development of uncoiled shells, which reaches its 
climax in the lower and middle Cretaceous, is considered 
as a sign of degeneration by many palaeontologists. To 
this view I cannot assent. The best instance of this mode 
of development is Lytoceras. It is one of the most con- 
servative types from the Lias up to the Neocomian, when 
suddenly a large number of uncoiled, straight, hook- 
shaped and even turriliticonic genera branch off from the 
old stem. All this stock flourishes for a considerable 
time. This does not mean degeneration, but adaptation 
to new and different forms of marine life, from a ben- 
thonic swimming to a chiefly creeping or even sessile 
mode of living. Still less can Lytoceras be stigmatized 
as degenerating, if we take into consideration the fact 
that it survives all its uncoiled offspring and persists 
into the Senonian stage. I consequently find in the devel- 
opment of uncoiled shapes in the family Ceratitidae, which 
begins with Choristoceras in the upper Carnic and reaches 
its climax in the lower Noric substage, a sign of increased 
adaptability to new modes of life, not of degeneration. 

Peculiarities of Constitution. — It is only touching the 
fringe of a great subject, if I venture to call attention to 
Brocchi's hypothesis, that the gradual and successive 
disappearance of species might be regulated by a constant 
law, that their death, like that of individuals, might 
depend on certain peculiarities of constitution. In our 
special case, as in many others, the extermination of a 
large and flourishing group of animals can be explained 
satisfactorily neither by external nor by such internal 
causes as are accessible to examination. In such cases 
Brocchi's hypothesis, although dealing with powers and 
influences of a still hitherto obscure nature, may yet serve 
its purpose as a first attempt to approach the solution of 
a hitherto unexplained problem. 

University of Vienna. 

Berry — Saccoglottis, Recent and Fossil. 127 

Art. XI. — Saccoglottis, Recent and Fossil; by 
Edward W. Berry. 

One of the most interesting of the romantic assemblages 
of fruits and seeds that constitute the sea drift, typically 
developed in the tropics, is Saccoglottis amazonica. 
Although the plant itself was described by Martius from 
the lower Amazon, its strange fruits had been known in 
Europe for over two centuries before their identity 
became known. They were figured by Clusius in 1605 
and mentioned by Sloane in 1696 but it was not until 1889 
that their botanical identification was accomplished, the 
details of this story having been told by both Morris 1 
and Guppy. 2 

The latter author deals very fully with this species, 
whose fruits are widely distributed in the Antillean sea 
drift, and are occasionally washed ashore in Europe. 
The fruits possess great buoyancy because of the ligneous 
pericarp and the numerous large resin cysts which it 
contains. Although the fruits are such ideal ocean 
travellers there is no evidence that they have established 
themselves on any of the Antilles where they are habit- 
ually washed ashore, unless the few plants in southern 
Trinidad have been introduced in this way through the 
agency of the Orinoco drift. 

According to Guppy this species is an inhabitant of 
the estuarine forests of the great rivers of Brazil, the 
Guianas, and Venezuela. As far as I know it has never 
been recorded from Colombia or Central America or 
anywhere on the Pacific coast. Great interest, therefore, 
attaches to my finding the fruits in 1919 near Old Panama 
in the sea drift of Panama Bay. Obviously the parent 
plant must grow somewhere on the Pacific watershed of 
Central or Northern South America. Two of these fruits 
from Panama Bay are shown in the accompanying figures. 
One with the warty sarcotesta intact and indicative of the 
fruit not having been in the water a long time, and the 
other with the outer coat worn away thus exposing the 
resin cavities and representing the usual form of preser- 
vation of these fruits in the Antillean sea drift. 

1 Morris, D., Nature, Jan. 31, 1889; Nov. 21, 1895. 

- Guppy, H. B. ; Plants, Seeds and Currents in the West Indies and 
Azores, pp. 133-137, 1917. 

Am. Jour. Sci.— Fifth Series. Vol. IV, No. 20 —August, 1922. 

128 Berry — Sacco glottis, Recent and Fossil. 

The genus Saccoglottis consists of about ten existing 
species found in the region from Brazil to Venezuela, and 
is a member of the restricted family Humiriaceae of the 
order Geraniales. The family is usually divided into the 
three genera Humiria, Vantanea, and Saccoglottis, and 
all of the known species are dwellers in the wet forests 

Fig. 1. 

Explanation of Fig. 1. 

a. Saccoglottis tertiaria Berry. Side view of a prolate form. 

o. End view of same showing the 5 seeds. 

c. A single seed of same. 

cl. Saccoglottis amazonica Martius from Panama Bay. 

e. End view of a worn fruit of this species showing seeds. 

/. Same in side view. 

of Brazil, the Guianas, Venezuela and eastern Peru, 
except for the single species Saccoglottis or Humiria 
gabonensis Urban which is sometimes considered the type 
of a fourth genus — Aubrya. 

The presence of Saccoglottis on the Pacific coast would 
suggest that the genus was an inhabitant of this general 

Berry — Saccoglottis, Recent and Fossil. 129 

region before the Isthmus of Panama was closed, and the 
presence of a well marked fossil species, to be described 
presently, suggests an American origin for the family, 
and suggests further, that the single west African coastal 
species reached that continent either by means of an 
equatorial counter current, or before the continental out- 
lines had assumed their modern form. As I have pointed 
out on a former occasion, there are a number of facts 
which suggest that Guppy, in his admirable studies on 
distribution, has underestimated the possibilities in this 
direction. Although the main equatorial currents might 
be expected to carry coastal types with seaworthy fruits 
from the Old to the New World, I see slight evidence of 
this ever having taken place, and there is a considerable 
body of evidence of dispersal in the opposite direction. 
If the present currents in the equatorial Atlantic pre- 
clude effective dispersal from west to east then we are 
forced to assume that the Tertiary oceanic circulation in 
this region differed from its present arrangement, and 
this could readily be brought about by changed conti- 
nental outlines, even though we are not in a position to 
predict at the present time just what these outlines were. 
One of the most interesting localities that I visited in 
1919 was a place called Pisllypampa in the mountains 
north of Cochabamba, Bolivia. Here on a bleak and 
treeless pampa at an elevation of 11,800 feet I found a 
rich tropical flora of Pliocene age preserved in beds of 
tuff. This flora will be described in full in the Hopkins 
Studies in Geology. One of the most abundant elements 
in this Pliocene flora were the fruits of a species of Sac- 
coglottis, which may be described as follows : 

Saccoglottis tertiaria Berry, n. sp. 

Fruits relatively small, varying from globular to pro- 
late spheroidal in form, sometimes somewhat flattened by 
pressure during fossilization. Drupaceous in character, 
the thin outer flesh (sarcotesta or epicarp) being pre- 
served as a carbonaceous incrustation in several speci- 
mens. The bulk of the fruit consisting of a woody stone 
or pericarp. The surface is slightly irregularly mam- 
nlilated or warty, and thickly impregnated with resinous 
cysts, whose cavities conspicuously and thickly pit the 
surface of the woody stone with depressions from 1 to 2 

130 Berry — Saccoglottis, Recent and Fossil. 

millimeters in diameter. The stone has imbedded in it 
five large seeds, arranged symmetrically around the 
central axis, and these appear to break away tardily on 
drying as in Saccoglottis amazonica Martins, since several 
are found as fossils in a detached state. These seeds are 
narrowly elliptical in surface outline, about 2 centimeters 
long and 7 millimeters wide ; the inner margins are trun- 
cated to a central, nearly straight, gable-like keel; the 
thickness of the seed, i.e., measured radially with respect 
to the fruit as a whole, being about 6 millimeters. There 
is some slight inequality in the development of the indi- 
vidual seeds, but generally all five seeds are nearly equally 
developed. The total dimensions of the fruit are from 
2 to 2.25 centimeters in length and from 1.6 to 1.9 centi- 
meters in diameter. No leaves that could be correlated 
with these fruits were found in the deposits. 

No plant family except the Humiriaceae has the features 
shown by these fruits — thin flesh, woody stone with 
numerous resin cysts, and 5 radially symmetrically 
arranged large seeds. I am unable to state the nearest 
living relative of these fruits, being much hampered by 
the lack of comparative material in the larger herbaria, 
and this is well illustrated by the number of years that 
elapsed before the botanists at Kew succeeded in deter- 
mining the fruits of Saccoglottis amazonica. The fossil 
fruits, as may be seen in the accompanying illustrations, 
are in a good state of preservation. They are certainly 
referable to the Humiriaceae and are strikingly like those 
of Saccoglottis amazonica, but as just stated, I have not 
seen the fruits of the majority of the Humiriaceae. I 
have ventured to refer the fossil to Saccoglottis which it 
so much resembles, and in any event it affords a striking 
glimpse into the past history of a family which is other- 
wise scarcely known in the geological record. 

Round — Crossotheca from R. I. Carboniferous. 131 

Art. XII. — A Crossotheca from the Rhode Island Car- 
boniferous; by Eda M. Eouxd. 

The presence of any form of fruiting body in the Penn- 
sylvanian is of interest, especially if it sheds light upon 
the evolution of seed plants. Among fossils of this class 
from the Ehocle Island coal basin have been found speci- 
mens of a new species of Crossotheca named from its 
diminutive size nana and described as follows: 

Crossotheca nana,' n. sp. — Stirps 5 mm. latus, paniculi laxi 
racemorum fructuum sustinere videtur. Paniculi for minimum 
5-6 em. longi vel longiores probabiliter ; secundaria^ divisiones 
ovatas fusiformas fruges sustinent, unaquaque 3-4 mm. longa, 
1-1% mm. lataque est. Fruges maturiores inclives sunt arcuta- 
tim paullum esse, ex raceme axe super quern sustinentur. Fruges 
lougitudinales in partes tres dividuntur. Partes ulteriores novi- 
ter divisae sunt et compositae oblongarum antheridium quae 
separatae vel ex parte conjunctae centrali parti videtur esse 
similis fimbriis. (Figs. 1, 2.) 

Stem 5 mm. wide, bearing loose panicles of clustered fruits. 
Panicles at least 5-6 em. long, probably longer; secondary divi- 
sions bearing oval fusiform fruits each 3-4 mm. long by I-IV2 mm - 
in diameter, in more mature forms inclined to be slightly arcuate, 
the more convex side being farthest from the axis of the raceme 
on which they are borne. The fruits appear to be divided longi- 
tudinally into three parts. The outer portions, which are again 
divided, are composed of oblong antheridia which are either 
separate or partly attached to the central part like fringe. 

In seeking for evidences of the presence of Crossotheca 
nana in other parts of the world, sketches by Gutbier of 
specimens 1 from the Zwickau coal basin of Saxony may 
be cited. The raceme figured by him (see fig. 3) is sug- 
gestive of the general appearance of the Rhode Island 
fossil when viewed under low magnification. Gutbier, 
however, makes no reference to his figures but in another 
text speaks of them as Sphenopteris allosanrioides, 2 a 
name which gives an interesting sidelight upon his inter- 
pretation of the fossil and shows that he regarded it as 
the fertile portion of a fern closely allied to the modern 
cliff brake. 

1 Gutbier, A. von: Abdriicke u. Verstein. Zwickauer, Atlas, PI. 10, figs. 
4-4b, 1836. 

2 Keichenbach, Ludwig: in Gaea von Sachsen (Verst. Uebersachsen) 1843. 

132 Round — Crossotheca from R. I. Carboniferous. 

Fig. i. 

Fig. 1. — Crossotheca nana Round, nat. size, Pawtucket, R. I. (No. 1012 
Roger Williams Park collection, Prov., R. I.) 

Fig. 2. 


Fig. 2. — Crossotheca nana; detail, V 5, 

Bound — Crossotheca from R. I. Carboniferous. 133 

According to Kidston, 3 Crossotheca fruits are not, as 
formerly supposed, the exannulate fertile pinnae of Pter- 
idophytes allied to the Marattiaceae but represent micro- 
sporangiate parts of a Pteridosperm, the sterile forms of 
which are of the Sphenopterid or Pecopterid type. 

Fig. 3. — Showing specimen as figured by Gutbier. 

Three typical Crossotheca species from the European 
coal flora show noteworthy differences from the American 
form. Crossotheca schatzlarensis Stur 4 consists of a 
more complex panicle than the Rhode Island specimen, 
the units of which are divided into four to eight antheridia 
as contrasted with fifteen to twenty in Crossotheca nana. 
The size of each unit, however, is about the same as that 
of the Rhode Island form although the proportions are 
very different, being loose where Crossotheca nana is 
compact. Sphenopteris (Crossotheca) Crepini Zeiller 5 
is about the size of the Rhode Island species but more 
stout in form and simple in details. Sphenopteris (Cros- 

3 Kidston, E. : Phil. Trans. Eoy. Soc, vol. 198B, pp. 413-445, Pis. 25-28, 
1908. Les vegetaux houillers recueillis dans le Hainaut beige, p. 41, 1909. 

4 Kidston, K. : Proc. Eoy. Phys. Soc, vol. 9, Pi. 21, figs. 1-6, 1888. 

5 Zeiller, E. : Bassin houiller de A r alenciennes, PI. 13, figs. 1-3, 1886. 

134: Round — Crossotheca from R. I. Carboniferous. 



















e=- S 


ig 1 
o .a 

^ 2 

O M 

r2 ^ 
^ O 


J2 S 

^ | 

03 i> 

03 f> 






•tnpijaq^ire jo qjpiM £ ^ 

£ CO 

•Brpt.i9q;nB jo q}§uarj g" 

s.iaqumn ^m.ij jo 
apis ipi3a 'Bipiiaq^ny 

•^iiuj jo iripiM. ' £ <m 

•iniXi io qiSuaq; ' g "=? 

teia^x jo i^PTM 


p.iajBi jo qjSiiai: d 

uia^s mBiu 03. araaDB.i 
TBia^iq 'uoixayuj 

•aiuaa^.i jo q^pi/VV 


■auiaDB.i jo q;.Snaq; ^ 











] | 




















«pi i3 

C3 r— 

=0 . g 

cr o oc 
1 °° s 

T3 ^ ^3 
CO? § 

cS°> s 

. bjc . 

=4H CC «H 

* 00 -t- 

Round — Crossotheca from R. I. Carboniferous. 135 

sotheca) B onlay i Zeiller, 6 however, has about the same 
proportions as Crossotheca nana but is about four times 
larger. Details of comparison of all these species are 
shown in the accompanying table. 

Specimens of Crossotheca nana have been found in the 
Pawtucket and Portsmouth sections of Rhode Island, an 
indication that the species was somewhat widespread in 
the Narragansett Basin during the Carboniferous. 

Paleontological Laboratory, 
Brown University. 

6 Ibid., PI. 4, fig. 4. 

136 F. H. Knowlton — Fossil Dogivood Flower. 

Aet. XIII. — A Fossil Dogivood Flower; by F. H. 


Fossil flowers are of such rare occurrence in this coun- 
try that when one conies to light it seems to merit an early 
description. In working up some material from the Fort 
Union formation (Eocene) in the Glenrock coal field, Con- 
verse County, Wyoming, I found the specimen here 
described, which is so obviously a dogwood "flower" that 
there is no hesitation in referring it to Cornus. It may 
be named Cornus speciosissima and described as follows : 

Involucral bracts 4, closely sessile, elliptical or elliptical- 
obovate in shape, rather obtuse at the apex where the tip is thick- 
ened. The bracts appear to have been rather thick and have the 
veins strong and all converging in the thickened tip. The length 
of the bracts is 18 or 20 millimeters and their greatest width 12 
or 13 millimeters. The spread of the perfect ' ' flower ' ' must have 
been nearly 4 centimeters. 

Although no evidence of the peduncle can now be 
detected it is clearly the under side of the whorl of invo- 
lucral bracts that is exposed. At the base of the best 
preserved one of the four bracts the surface is seen to be 
finely striate and somewhat wrinkled, and showing 
through are five or six dark circles which undoubtedly 
represent a part of the cluster of flowers on the upper 
side. This completes the evidence necessary for its 
reference to Cornus. 

Cornus has some forty or fifty living species widely 
distributed over the three continents of the northern 
hemisphere, with a single species crossing the equator 
and reaching Peru. The genus is sharply separated into 
several groups which have sometimes been given separate 
generic rank, but it seems to me that they are all best 
retained under Cornus. In one group which embraces 
the majority of the species the flowers are cymose and 
not involucrate, while in another group the flowers are 
capitate with an involucre of large usually white bracts. 

Cornus speciosissima belongs, of course, to the invo- 
lucrate group and as nearly as can be made out seems 
to be most like Cornus canadensis Linne, the dwarf cornel 
or bunch-berry, which ranges from Newfoundland to 

F. H. Knoidton — Fossil Dogwood Flower. 137 

Alaska and south to New Jersey, Ohio, Colorado, and 
California. This species is herbaceous above and woody 
at the base, with a whorl of leaves at the top, and long- 
stalked flowers. The fossil species under discussion is 
slightly larger than C. canadensis, has the bracts more 
nearly elliptical than obovate, and seemingly thicker in 
texture. No leaves of Cornus were found in direct asso- 
ciation with this specimen, but the collection was a small 
one and so such leaves might have escaped observation. 




Figure 1. — Cornus speciosissima, new species, showing the four involucral 

Figure 2. — Bract showing thickened tip. 

There are, however, two well-defined species in the Fort 
Union, based on leaves, namely Cornus newberryi Hollick 
and Cornus fosteri Ward. Judging from the size and 
texture of the leaves of these forms they were shrubby 
species and not of the herbaceous type of Cornus cana- 
densis. Without being in any way positive about it, it 
seems probable that Cornus speciosissima was borne on 
a plant of the shrubby type, and not by one of the herba- 
ceous type, although its flowers do resemble that of C. 

138 F. H. Knowlton — Fossil Dogwood Flower. 

Over twenty fossil species of Cornus have been 
described from North America, all based on leaves. 
These range in age from Middle Cretaceous to Pleisto- 
cene. In the Old World over forty fossil species have 
been named, two of which — both from the Miocene of 
Switzerland — are founded on the involucral bracts. Of 
these, Cornus bilchii Heer 1 has the bracts oblong in shape 
but not conspicuously thickened at the tip, and C. apicu- 
lata Heer, 2 with a long, slender evidently hardened point. 
Both these species are based on small detached bracts, 
and while C. speciosissima agrees closest in shape with C. 
bilchii, they are distinct. Cornus apiculata is wholly 
unlike C. speciosissima. 

The exact locality whence Cornus speciosissima came 
is the west bank of Cole Creek, about 1 mile east of Big 
Muddy, Converse County, Wyo. (Sec. 36, T. 33 N., R. 
77 W.). Collected by John B. Eeeside, Jr., September 5, 
1913. The types are preserved in the United States 
National Museum, Nos. 36616, 36617. 

U. S. Geological Survey, 
Washington, D. C. 

1 Flora fossilis helvetise, vol. 2, p. 27, pi. CV, figs. 6, 7, 1859. 

2 Idem, p. 28, pi. CV, figs. 10, 11. 

A. Wandke — Intrusive Rocks. 

Art. XIV. — Intrusive Rocks of the Portsmouth Basin, 
Maine and New Hampshire; by Alfred Wandke, 
Foxcroft House, Cambridge, Mass. 


Location. — The portion of the Portsmouth Basin to be 
treated in this paper includes 500 square miles of terri- 
tory that lies partly in the southwestern corner of Maine, 
and partly in the southeastern corner of New Hampshire. 
The area under consideration is easily accessible by 
rail being served by the Eastern and Western Divisions 
of the Boston and Maine Railroad. The Atlantic Shore 
Street Railway passes through much of the country not 
touched by the steam road and thus but little remains 
that is not easy of access. 

The portion of this area in Maine, except for the imme- 
diate vicinity of the shore where a thriving business is 
done entertaining summer visitors, is sparsely settled; 
that in Xew Hampshire, traversed by several rivers upon 
which manufacturing industries have been established, 
although not densely populated, contains a number of 
prosperous communities. 

Field Work. — This paper is based upon field work 
done during the years 1915, 1916 and 1917. During this 
time members of the United States Geological Survey 
were conducting an investigation of the geology of 
southwestern Maine and conferences with Dr. L. Laforge 
of the Survey aided greatly in deciphering the obscure 

Previous Work. — Except for the work of Jackson 1 of 
the Maine Survey in 1839, and of Hitchcock 2 of the New 
Hampshire Survey in 1868, but little had been clone in 
this field. At various intervals since the publication 
of the reconnaissances of these two men, notes mention- 
ing the area have appeared in several publications : the 
clays of South Berwick, Maine, which contain Pleistocene 
fossils, have been cited in papers dealing with the 
elevation of the coast of Maine ; Kemp 3 in 1890 described 
some of the dikes at Kennebunkport and Bald Head Cliffs, 

1 Geology of Maine, Augusta, Maine, 1S39. 

2 Geology of Xew Hampshire, Concord, Xew Hampshire, 1878. 

3 Amer. Geologist, vol. 5, 1890. 

140 A. Wan dice — Intrusive Rocks of 

Maine, and mentioned similar occurrences at Portsmouth, 
New Hampshire ; G. P. Merrill 4 noted the Durham granite 
in his i i Building Stones ' ' ; Dale 5 briefly referred to the 
Berwick "black granite"; Powers 6 in 1914 had occasion 
to visit Bald Head Cliffs while collecting material for his 
paper on the inclusions in dikes. Recently the U. S. 
Geological Survey began a systematic study of this 
region and two papers 7 relating to the geology have 

Acknowledgments. — In connection with this paper 
acknowledgments are due to Professor R. A. Daly of 
Harvard University under whose guidance the area was 
studied; to Professors John E. Wolff and Charles 
Palaclie, both of Harvard, for helpful aid in the micro- 
scopic investigation; to Dr. L. Laforge and Dr. Frank 
Katz of the IT. S. Geological Survey for suggestions 
regarding the characteristics of the formations. 


The Portsmouth Basin lies within the tilted and dis- 
sected Cretaceous peneplain of the Northern Appala- 
chian Geological Province. The portion of this pene- 
plain near the shore, spoken of as the coastal lowland, 
to which the area under consideration properly belongs, 
is a land surface of low relief and slight diversity. It 
rises gently toward the interior and its few hills, 
either monadnocks or morainal material left by the 
Pleistocene glaciers, rarely attain an elevation of more 
than 300 feet above sea level. Its valleys, except for 
those of the short coastal streams, are relatively broad 
and shallow depressions whose courses, determined 
largely during post-Cretaceous time, were but slightly 
affected by the glaciers of the Pleistocene period. The 
shore line is marked by a succession of bold rocky head- 
lands between which are marshes fronted by excellent 
sand beaches. Wherever rivers enter the ocean they 
generally afford good harbors, this being one result of 
a recent drowning of the shore. 

4 G. P. Merrill, Stones for Building, 1897. 

5 T. N. Dale, U. S. Geol. Survey, Bull. 313, 1907. 

6 S. Powers, Jour, of Geol., vol. 23, 1915. 

7 U. S. Geol. Survey, Prof. Paper 108, 1917. 

Portsmouth Basin, Me. and N. H. 1-tl 

Areal Geology. 

General Statement. — The Portsmouth Basin resembles 
in many respects the other fragmentary geological basins 
which have been recognized in the as yet imperfectly 
known area of the Northern Appalachians. The rocks 
consist of fine-grained steeply inclined sediments 
supposed to be of Upper Carboniferous age, of both intru- 
sive and effusive igneous rocks, and of some thoroughly 
metamorphosed rocks of doubtful origin and unknown 
age which have been classified as gneisses and schists. 
The sediments and the metamorphic rocks have a general 
strike of X 45° E, but local departures from this general 
direction are numerous ; the dikes of the region follow 
a northeasterly trend; and the batholiths are elongated 
in the same direction. The dip of the stratification of 
the sediments and of the schistosity of the metamorphics 
is, on the whole, to the northwest, but frequent departures 
from this direction indicate, perhaps, local folds and 
crumplings. The gneisses and schists occur both to the 
northwest and to the southeast of the sediments of Upper 
Carboniferous age, and since the highly metamorphosed 
rocks almost close in on the sediments to the southwest, 
a basin-like arrangement results. 

Sedimentary Rocks. 

General Statement. — Hitchcock s in his survey made no 
attempt to separate the rocks of the Portsmouth Basin 
into groups or formations, as is indicated by the following 
quotation: "This name (Merrimack Group) was infor- 
mally applied by my father to the mica schists, slates, and 
quartzites contained in the Merrimack River, in Massa- 
chusetts. They skirt the Exeter sienites in New Hamp- 
shire, lying in troughs in an anticlinal. They probably 
belong to the earliest Silurian." On purely lithologic 
grounds it seems possible to separate this "Merrimack 
group of Hitchcock's into two groups. The first contains 
three formations which have been called the Gonic schist a , 
the Berwick gneiss' 3 , and the Rye gneiss ; the second con- 

s Geology of New Hampshire, Concord, Xew Hampshire, p. 27, 1878. 
a Correlated by Katz, as a part of the Eliot formation but lithologically 

b Is the same as the Berwick gneiss of Katz. 
c Is the Algonkian complex of Katz. 

142 A. Wandke — Intrusive Rocks of 

sists of two formations, the Kittery quartzite d and the 
Eliot phyllite 6 . No definite age has been assigned to the 
first group nor can correlations be made between the 
different members ; the second group, it seems, can be 
traced to Worcester, Massachusetts, where fossil-bearing 
rocks, 9 Upper Carboniferous in age, have been found. 
The writer, however, has been able to trace the Kittery 
formation only as far as Lowell, Mass. From there on 
to Worcester, Mass., a distance of 30 miles, glacial 
material makes it difficult to follow the strata. The 
rocks at Worcester from which the fossils have been 
reported are lithologically unlike those of the Portsmouth 
Basin. In the hand specimen the typical Kittery quartz- 
ite is almost identical in appearance with the Isleboro 10 
formation of Rockland, Maine, perhaps of Cambrian age. 
It would seem, therefore, that until fossils are found 
within the Kittery formation its age must remain a 
matter of inference. The Kittery and Eliot formations 
are however provisionally dated as Carboniferous, follow- 
ing the lead of the U. S. Geological Survey. 

Metamorphic Rocks of Unknown Age. 

General Statement. — This classification includes the 
Gonic schist and the Berwick gneiss which form the north- 
western boundary of the basin, and the Rye gneiss on the 
southeastern side. Limits have been assigned to the 
extent of these formations with the understanding that 
they are not to be regarded as fixed. 

Gonic Schist. — The Gonic schist has been named from 
the typical exposures occurring in the falls of the Cocheco 
River at Gonic, New Hampshire. The rock is a biotite- 
garnet schist evidently produced by dynamic and 
contact metamorphism of argillaceous and arenaceous 
sediments. Cutting across the schistosity of the forma- 
tion, although occasionally paralleling the same, are a 
great many stringers of coarse pegmatite and veins of 
quartz. It is entirely probable that these stringers and 
veins are genetically related to the intrusive pegmatitic 

d The same as the Katz 's Kittery quartzite. 
e Is the Eliot slate of Katz. 

9 David White, Jour. Wash. Acad. Sci., p. 115, 1914. 

10 U. S. Geol. Survey, Folio 158, 1908. 

Portsmouth Basin, Me. and N. H. 143 

granite which outcrops less than 2000 feet distant from 
the exposures of Gonic schist. The formation has, in 
general, a gentle northwesterly dip, but departures from 
the normal dip and strike are abundant. The thickness 
of the formation is unknown. 

Berwick Gneiss. — This formation is typically devel- 
oped at the falls of the Salmon Falls River in Berwick, 
Maine. Dynamic and static metamorphism have entirely 
obscured the original texture and composition of the 
rocks which now consist of thin bands of mica schist 
alternating with well banded paragneiss. The formation 
appears to have been derived from an argillaceous sand- 
stone or a graywacke. Because of metamorphism some 
of the bands are now characterized by giaucophane. 
Associated with the amphibole are feldspar, quartz, 
biotite, chlorite, pyrite, and titanite. Quartz veins vary- 
ing in width from mere stringers up to masses two feet 
in thickness frequently cut the formation. Although 
some of the veins are developed parallel to the foliation 
of the gneiss, the general tendency is for them to occur 
in cross-cutting relationships. The thickness could not 
be determined. Departures from the general north- 
westerly dip of the foliation, which in rather limited 
areas may undergo wide variations, suggest that the 
formation is compressed into several tight folds which 
have been overturned to the southeast. Little can be 
said of the position of this formation in relation to its 

Rye Gneiss. — The Rye gneiss has been so called from 
the typical exposures that occur along the Rye coast of 
New Hampshire. The most northerly outcrops of rock 
belonging to this formation are found on Gerrish's 
Island, Kittery, Maine. From this locality the rocks 
have been traced south and southwest into Portsmouth, 
Rye, North Hampton and Hampton, New Hampshire. 
At Hampton Falls the country becomes drift-covered 
and for a space of six miles westward across the towns 
of Kensington and East Kensington no outcrops are to 
be seen. In the town of Kingston gneisses outcrop at 
Rock Rimmon Hill. 

At the type-locality the gneiss is well banded, consist- 
ing of alternations of light feldspathic layers, which 
become pegmatitic with dark-fine biotite-rich bands that 

A.m. Jour. Sci.— Fifth Series, Vol. IV, No. 20.— August, 1922. 

1-M A. Wandke — Intrusive Rocks of 

are schistose. On Gerrish's Island the rock is fine- 
grained, the banding characteristic of the typical gneiss 
being absent. As narrow stringers of feldspathic mate- 
rial penetrate the rock it loses its sedimentary habit. 
The feldspathic stringers may increase in size until 
they attain a width of fifteen feet. The entire occur- 
rence is suggestive of the feldspathization described by 
Wherry 11 in the pre- Cambrian highlands of eastern 
Pennsylvania. On Newcastle Island the process of 
feldspathization is well marked, and the gneisses become 
distinctly granitized. 

A number of dikes cut the gneisses, but in not a single 
instance in the good exposures along the coast has the 
metamorphism so characteristic of the gneisses affected 
a dike. Since some of these dikes are earlier than the 
batholiths (to be described later) the period of graniti- 
zation and feldspathization must be earlier than the 
period of batholitic intrusion. 

Grouped with these gneisses are some rocks which sug- 
gest altered basic volcanic flows. These have not been 
studied in detail. 

Sedimentary Bocks Upper Carboniferous in Age. 

General Statement. — The Upper Carboniferous rocks 
form a single group consisting of two formations. The 
Kittery quartzite is the older of these and the Eliot 
phyllite the younger. Boundary lines between the two 
cannot be drawn with any great degree of accuracy, 
and the lack of exposures and the prevalent glacial drift 
prevent a close correlation. The two formations appear 
to be conformable. 

Kittery Quartzite. — The typical Kittery quartzite is 
found just north of the bridge of the York Harbor and 
Beach Railroad that crosses Spruce Creek in the town 
of Kittery, Maine. At this locality the formation con- 
sists of a series of thin beds of red phyllite that alternate 
with thicker strata of a dense fine-grained grayish green 
or bluish quartzite. Variations from the normal are, 
however, abundant, the differences being due in part to 
primary causes during the time of deposition and in part 
to secondary changes during and after the period of fold- 

11 Paper read before the Geol. Soe. of America, 1916. 

Portsmouth Basin, Me. and N. H. 145 

ing. The thickness cannot be accurately stated. In a 
section from Godfrey's Cove northwest to the contact of 
the sediments with the granites, a distance of eight miles, 
the rocks dip almost invariably steeply to the northwest, 
and although several folds seem to be indicated, the 
covering of glacial drift obscures the structure. Along 
the coast where outcrops are more numerous a duplica- 
tion of lithologically similar rocks appears at such inter- 
vals as to suggest that the formation has a thickness of 
about 2000 feet. 

The lower limit of the Kittery could not be established : 
in the southeastern part of the area the formation is 
seemingly a faulted contact with the Rye gneiss ; in the 
northwestern part it appears to stand in a faulted 
relationship to the Berwick gneiss. Multitudes of dikes 
cut this formation. 

Eliot Phyllite. — This phyllite is typically developed 
in the town of Eliot, Maine. It is, as a rule, gray in 
color but red, brown, black, and buff phases are common. 
Slight variations in texture and composition are observed 
and in the formation calcareous and carbonaceous phyl- 
lites, true slates, and a light yellow to brown gray 
crumpled and easily eroded argillite have been included. 
Quartz veins and dikes are frequently seen cutting the 
formation. Because of the lack of exposures it becomes 
difficult to establish a standard section, and hence the 
thickness can at best merely be estimated. South of 
Eliot where exposures are most numerous the thickness 
would seem not to exceed 2500 feet. Since the passage 
from quartzite to phyllite is effected by transitional beds 
which gradually change from quartzite with thin beds 
of phyllite to homogeneous phyllite or shale, and because 
the prevalent dips are such as to carry the quartzite 
constantly beneath the shale or phyllite, it seems certain 
that the Eliot and Kittery formations are conformable 
and that the Eliot is the younger. 

Igneous Rocks. 

General Statement. — The igneous rocks of the area 
include a great many dikes, several subjacent bodies, 
a few effusives, and some mixed gneisses. The mixed 
gneisses are pre-Carboniferous in age and occur in the 

116 A. Wandke — Intrusive Rocks of 

Rye formation. The effusives, or at least rocks resem- 
bling metamorphosed effusives, occur in a narrow belt 
at Portsmouth, New Hampshire. No attempt was made 
to study them. The discussion will center upon the dikes 
and subjacent bodies which for the most part cut the 
Upper Carboniferous rocks. 


The dikes, because of their number, their wide range 
in composition, and the manner and order of their 
intrusion form a group of rocks well deserving of more 
detailed study. Although every outcrop carries one or 
more of them, the best display of dikes occurs along the 
coastal section from Perkin's Cove to Brave Boat 
Harbor. In these eleven miles there are several hundred, 
perhaps a thousand, dikes and they vary in width from 
stringers hardly thicker than the blade of a case knife 
up to 200 feet. Multiple and composite examples 
abound. In several instances three or more dikes of 
the same or of contrasted compositions may occupy a 
single fissure. In composition they range from an 
olivine-bearing lamprophyre through diabase, diorite, 
and granite porphyry to paisanite and aplite. Some 
^abound in inclusions, which being fragments either 
brought up from depth or torn from the adjacent walls, 
occur in various stages of assimilation. In short, the 
dikes of the area form a remarkable display and illustrate 
most of the phenomena associated with intrusions of 
this type. All of the observed dikes are later than both 
the period of folding and the period of granitization 
which produced the Rye gneisses, and for the most part 
appear younger than the quartz veins and the metamor- 
phism which characterize the outcrops from Perkin's 
Cove to Brave Boat Harbor. In a broad way they may 
be separated into two groups, 12 (a) those earlier, and (b), 
those later than the stocks and batholiths. On the basis 
of composition each of these groups may be subdivided 
and such a sub-division is helpful in gaining a clearer 
conception of the changes in the magma from which the 
dikes were derived. 

12 A similar grouping was made by C. H. Clapp for the dike rocks of 
Essex County, Mass., U. S. Geol. Surv., Bull. 704, p. 107, 1921. 

Portsmouth Basin, Me. and N. H. 147 

First Group. 

Although this group has been divided into three sub- 
groups, — diabases, diorites, and granite porphyries, no 
hard and fast line of division can be drawn, since the dia- 
bases grade into the diorites and the diorites show transi- 
tion phases which suggest the granite porplryries. 

Diabases. — For the most part the diabase dikes are the 
oldest of the region. In the batholiths other diabasic 
dikes are seen which indicate a second period of diabase 
intrusion as a late phase of igneous activity. This 
region illustrates therefore the initiation of a period of 
intrusion by a great development of diabase dikes and 
the recurrence of similar rocks at the close of the 
irruptive period. 

Diorites. — As the rocks in the preceding class grow 
lighter in color they grade into the diorites. The lighter 
colored members of the group have a banded appearance 
due to the development of innumerable segregations of 
quartz and feldspar in subparallel lines that follow the 
contacts. The development of quartz in these rocks is 
worthy of note for it may indicate that the magma of 
which they are offshoots began to approach a quartz 
diorite in composition. 

Granite Porphyries. — Under this heading have been 
placed all of the light-colored typically porphyritic dike 
rocks of which the phenocrysts are invariably quartz 
and feldspar. Some of these are characterized by 
resorbed quartz phenocrysts, by pyrrhotite instead of 
pyrite, and by a groundmass which consists almost 
entirely of myrmekite. The myrmekite may well have 
resulted from the escape of volatile components as has 
been suggested by Sederholm. 13 

Second Group . 

General Statement. — This group consists of two sub- 
groups of diaschistic dikes, — the one contains the paisan- 
ites, tinguaites, and camptonites; the other the aplites 
and the late diabases. The first sub-group is gener- 
ally associated with rocks that belong to the alka- 
line clan; the second with those which belong to the sub- 
alkaline clan. 

13 Bulletin de la Commission Geologique de Finland, No. 48, p. 81, 1916. 

14S A. Wandke— Intrusive Rocks of 

- First Sub-group. 

Tinguaites. — The tinguaite dikes vary from a deep 
blue to a light gray in color. The feldspar is anortho- 
clase. The aegerite needles which dominate the ground- 
mass give the rock its characteristic color. 

Camptonites. — The camptonites are poorly exposed, 
but wherever found they are characteristically developed. 
They are full of inclusions and contain large porphy- 
ritic crystals of glistening poikilitic hornblende. The 
hornblende is full of feldspar and apatite. 

Paisanite. — A single example of this type of dike was 
found near the crest of a hill just north of "Scotland," 
York, Maine. The minerals present are quartz, albite, 
microcline, microperthite, aegerite, riebeckite, arfved- 
sonite, zoisite, and an undetermined titanif erous mineral, 
brown in color and platy in habit. Both zoisite and the 
feldspars are poikilitically intergrown.. 

Second Sub-group. 

Aplites. — The aplites, present in each of the subjacent 
intrusives, are composed essentially of feldspar and 
quartz, the darker minerals being sparingly developed. 
Their composition and texture present no unusual 

Diabases. — The diabases of the second generation 
differ but little from those of the earlier period. They 
are found cross cutting all of the other types of rock. 

Stocks and Batholiths. 

General Statement— "Large intrusive bodies occupy a 
considerable part of the Portsmouth Basin. They are 
of especial interest because of their contrasted composi- 
tions, their contact actions, and their method of emplace- 
ment. For purposes of easy reference these bodies have 
been named the Rochester biotite granite; Durham 
quartz diorite; Hampton granodiorite ; Agamenticus 
complex, which consists of biotite granite, gabbro, sye- 
nite, and alkaline granite ; Cape Neddick gabbro ; York 
Harbor biotite granite; and the Brave Boat Harbor 
biotite granite. 

Portsmouth Basin, Me. and N. H. U9 

The New England province seems to have been affected 
by two periods 14 of batholithic intrusion. The first is 
usually dated as Devonian ( I) and is characterized by 
granites, granodiorites, and quartz diorites ; the second, 
or Carboniferous, is characterized by alkaline granites 
of which the Quincy granite is the typical example. It 
is probable that the subjacent rocks of the Portsmouth 
Basin fall into these two groups. In (1) may be placed 
the Rochester granite; in (2), the Durham, Hampton, 
Agamenticus, Cape Xeddick, and York Harbor occur- 
rences. The granite at Brave Boat Harbor is much more 
sheared than are the other nearby intrusives. It is 
doubtfully classified as a Carboniferous intrusive. 

Rochester Biotite-Granite. — This body has no unusual 
features of mineralogical composition. Outcrops are 
not numerous, the best exposures occurring in a few small 
quarries. The rock cuts the Gonic schist and may be a 
pre-Carboniferous intrusive. Pegmatite veins cut the 
granite and as a rule carry quartz, perthitic feldspar, 
and muscovite. 

Durham Quartz-Bio rite. — The Durham quartz diorite 
batholith is an elongated body that extends from two 
miles southwest of Exeter, X. H., to within one fourth 
mile of Eollingsford, X. H. It occupies about one third 
of the Dover quadrangle, has an extreme length of twenty 
miles, and a maximum width of four and one half miles. 
Outcrops showing the crosscutting relationships of the 
body are abundant and the boundaries as shoAvn on the 
map will need but slight revision. The elongation 
corresponds to the dominant strike of the invaded sedi- 
ments. In places inclusions of the Kittery formation 
occur in various stages of assimilation. 

Although the composition has been indicated as quartz 
diorite this term merely covers the dominant phase. A 
distinct gradation from a basic margin to an acidic 
interior is one of the features of this body. In a section 
taken across the batholith in a southeasterly direction 
through the town of Durham, the following variations 
in composition have been noted: a marginal phase of 
gabbro, quartz norite, quartz gabbro, quartz augite 
gabbro ; an intermediate phase of quartz augite diorite, 

14 B. K. Emerson, Geology of Massachusetts and Ehode Island, U. S. Geol. 
Survey, Bull. 597, p. 172, 1917. 

150 A. Wandke — Intrusive Rocks of 


J / <C'* I ? iZU'v? rz@ " MT AGAMENTICUS 

1 Berwick QuarhsLj)iorite v 

Portsmouth Basin, Me. and N. H. 151 

quartz cliorite, quartz biotite diorite ; and a central phase 
of granodiorite, granite, and granite aplite. 

Volumetric composition of two phases of the Durham 
quartz diorite : 


Plagioelase (Ab 3 An 7 — Ab 6 An 4 ) 63.15 

Mieroperthite to oligoclase 56.06 

Quartz 10.03 17.10 

Biotite 16.21 15.92 

Pvroxeue 8.91 .... 

Amphibole 1.68 10.10 

Apatite .01 

Totals 99.98 99.19 

I. Quartz diorite 300 feet from the contact. 
II. Granite from the central portion. 

Other bodies related to the quarts diorite. — Two other 
bodies within the area resemble the Durham batholith, 
not only in composition, but also in a similar gradation 
from basic margins to a more acidic central phase. One 
of these, a small stock, occurs about two miles east of 
North Berwick ; the other, also stock-like in habit, occurs 
at the intersection of the North. Hampton, Exeter and 
Newington boundaries. 

Cape Xeddiek Gabbro. — The Cape Xeddick gabbro 
forms a small oval stock measuring about one-half by 
three-fourths of a mile. It is well exposed and shows 
particularly fine contact phenomena as well as variations 
in composition. This stock shows four phases: (a) the 
contact phase, a medium to fine-grained rock rich in 
olivine, myrmekite, and apatite, and contains abundant 
inclusions of the invaded quartzites ; (b) a dark coarse- 
grained phase (phenocrysts may show a diameter up to 
2 cm.) having about equal amounts of the light and 
dark minerals which are poikilitically intergrown and 
characterized furthermore by a vertical banding parallel 
to the contacts; (c) a very dark, almost black phase, 
medium to coarse in grain, containing olivine embedded 
in large poikilitic crystals of hornblende, biotite, and 
titaniferous pyroxene; and (d) a light colored central 
phase composed largely of plagioelase feldspar accom- 
panied by small amounts of quartz and orthoclase. 

152 A. Wandke — Intrusive Rocks of 

Volumetric composition of the four phases of the Cape 
Neddick gabbro : 


Olivine 17.89 1.5.0 6.76 

Plagioclase 45.50 45.60 51.85 87.10 

Alkaline feldspar 1.17 .... .16 

Pyroxene 23.53 25.00 8.23 9.43 

Hornblende .... 10.01 21.42 .68 

Biotite 3.77 3.26 4.76 1.21 

Hypersthene 2.33 

Magnetite .. 3.62 13.82 6.79 1.55 

Apatite . 2.97 .... 

Totals . . . , 100.78 99.19 99.81 100.13 

I. Contact phase of the gabbro. 

II. Pyroxene-rich phase. 

III. Hornblende-rich phase. 

IV. Central feldspathic phase. 

Agamenticus Complex, 

This complex, so-called to signify that it consists of 
more than one kind of rock, is situated in York County, 
Maine, and has been named from Mount Agamenticus 
upon whose slopes three of the dominant rock types are 
to be found. 

Biotite Granite. — This rock forms the dominant phase 
of the batholith. The texture and composition vary from 
place to place, but on the whole the rock is light colored, 
has a slightly cataclastic structure, and a porphyritic 
habit. Near the Jewett station, York, Maine, a narrow 
vein carrying pyrite and molybdenum was found. This 
vein is the only evidence of ore mineralization seen in 
the entire area. 

Alkaline Granite. — This rock forms an irregular, 
roughly U-shaped body, the limb of the U enclosing 
biotite granite and the syenite stock. The typical granite 
is characterized by euhedral and anhedral perthitic 
feldspar, allotriomorphic to poikilitic arfvedsonite and 
aegerite-augite. Quartz-rich phases, aplitic tendencies 
and peculiar orbicular fine-grained contact phases are 
some of the departures from the normal. Although the 
usual amphibole is arfvedsonite, riebeckite is common; 
aegerite and aegerite-augite are the usual pyroxenes; 
ort'hite is seen in almost every section. 

Portsmouth Basin, Me. and X. H. 153 

Syenite. — The syenite forms an irregular stock-like 
mass measuring three and one-half by four miles. The 
field habit of the rock is varied, all gradations being seen 
from a fine-grained grayish-green quartz-free syenite to 
a coarsely pegmatitic rock consisting essentially of 
perthitic feldspar, quartz, and arfvedsonite. The oli- 
vine, fayalite, occurs rather abundantly in the contact 
phases ; in some of the central phases, arfvedsonite, 
riebeckite, orthite, aegerite, and zircon become locally 

Contacts of this rock against the alkaline granite are 
rather poor so that age relations cannot be established 
with certainty. Along the northwestern margin of the 
syenite this rock appears to dike the alkaline granite. 
The syenite, usually a quartz syenite (nordmarkite) in 
composition, is in turn cut by the lenticular masses of 
rather coarse-grained pegmat.ic quartz syenite, thus indi- 
cating an additional concentration of volatile components 
in the later stages of this intrusive. 

Volumetric composition of two phases of the Agamen- 
ticus complex : 


Alkaline feldspar 87.90 76.02 

Fayalite 3.76 

Aegerite augite 3.35 2.61 

Quartz 2.43 17.46 

Arfvedsonite 49 3.55 

Biotite 61 

Magnetite 73 .... 

Allanite .19 

Zircon .09 

Totals 99.27 99.92 

I. Border phase of the syenite. 
II. Typical alkaline granite from Mt. Agamenticus. 

. Berwick Quartz Diorite. — Apparently later than the 
main biotite granite is the small stock of quartz diorite 
situated about two miles east of Berwick. The contact 
phase is essentially a gabbro, but the central mass is a 
typical quartz diorite. This body may possibly be cor- 
related with the Cape Xeddick gabbro. 

Other Granitic Bodies. — In addition to the above men- 
tioned occurrences of granite there are other small bodies 

154 A. Wandke — Intrusive Rocks of 

such as those along the South side of York Harbor and 
along the northeastern side of Brave Boat Harbor. _ The 
Brave Boat Harbor granite is highly cataclastic. Neither 
body presents unusual characteristics. 

Emplacement of the Intrusives. 

Elongation. — The map shows that each of the larger 
intrusive bodies is distinctly elongated in a northeasterly 
direction, which corresponds in a general way to the 
dominant direction of folding for this region. In this 
respect they conform to one of the characteristics of 
batholiths as postulated by Daly, 15 that they should be 
elongated in the direction of the orogenic axes. 

Method of Emplacement. — The contacts of each of the 
large intrusive bodies have been traced as carefully as 
the exposures would admit. The mapping has shown 
that they are distinctly cross-cutting. Contact breccias 
are decidedly common and within each of the intrusive 
bodies are " islands'' of the invaded sediments whose 
strike and dip corresponds with that of the nearby 
country rock. These inliers are unquestionably roof 
pendants. The map on an earlier page will better show 
the relationships. The question of the isolation of the 
roof pendants becomes vital in considering the emplace- 
ment of the intrusives. Without going into the details 
as to how this emplacement was effected, the following 
table is offered in support of the method that seems to 
have played an important part : 

Specific Gravities } Q 

Temp, in Centi- Solid Molten 

grade 20 1000 1100 1200 1000 1100 1200 

Sediment 2.71 2.67 2.66 2.65 i 

Gabbro 2.92 2.85 2.84 2.83 2.67 2.66 2.65 

Quartz diorite. 2.81 2.77 2.76 2.75 2.58 2.57 2.56 

Alkaline granite 2.67 2.63 2.62 2.61 i 2.38 2.37 2.36 

The tabulated specific gravities indicate that a frag- 
ment of sedimentary rock similar to the Kittery quartz- 
ite in composition would tend to sink in a magma having 

15 Igneous Eocks and Their Origin, 1914. 

10 Day, Sosman and Hostetter, this Journal, vol. 37, 1914. 

Portsmouth Basin, Me. and N. H. 155 

the composition of a quartz diorite or an alkaline granite 
if the magma in question were molten at a temperature 
even less than 1000° C. The stoping hypothesis 17 may 
thus account for a large measure of the final emplacement, 

Although assimilation in place may have removed a 
part of the invaded sediments, and although block-fault- 
ing or differential lateral movement may account for a 
part of the width of these batholiths, nevertheless, the 
areas of contact breccias, the roof pendants, and the 
relative specific gravities of molten rock and engulfed 
fragments all seem to indicate that the intrusives of the 
Portsmouth Basin in their later stages have been 
emplaced to a large degree by magmatic stoping. 

Slightly different in its method of emplacement is the 
Cape Xeddick gabbro. Excellent exposures of this body 
are to be had, the northwestern contact being particu- 
larly illuminating. Here the waves have cut a marine 
bench about 100 feet wide and at low tide the rocks can be 
studied in detail. Basic and granitic dikes cut the Kit- 
tery quartzite which forms the country rock. Close to 
the intrusive the dikes and sediments are mixed by 
contact brecciation and contact metamorphism has almost 
obliterated the original lithologic differences. A short 
distance from the contact the sediments are seen to be 
crumpled, simple folds, overturned folds and overturns t 
faults all being present on a miniature scale. The crump- 
ling gives way 75 feet distant from the contact to gentle 
folding and this in turn grades into the characteristic 
steep dip and NE-SW strike of the sediments. The intru- 
sive was apparently emplaced in part by bodily forcing 
its containing walls apart and crumpling and mashing 
them. As shown by the peculiar quartz-rich contact 
phases a part of this shattered rock was assimilated; 
a part seems also to have been removed by being floated 
away. This latter condition is well illustrated by frozen- 
in isolated blocks of quartzite weighing several tons now 
removed a short distance from the contact. The figures 
dealing with specific gravities indicate that these blocks 
may have sunk in gabbroid magma, there to be slowly 
digested and assimilated at depth. 

17 For a detailed presentation of this hypothesis, see E. A. Daly, Igneous 
Eocks and Their Origin, p. 194, 1914. 

156 A. Wand he— Intrusive Rocks of 


The stocks and batholiths of the Portsmouth Basin, 
although not remarkable for an assemblage of rare rock 
types, nevertheless exhibit contrasted mineralogical com- 
positions. In the descriptions of the igneous bodies it 
was mentioned that the composition of the Durham 
quartz diorite varied from a basic marginal phase to an 
acidic central one ; that the Agamenticus complex seems 
to show a march of progress from calcic to alkalic types ; 
that the dikes in their cross-cutting, relationships indicate 
a change in composition of the parent magma from a 
basic to an acidic type. The foregoing statements appear 
to be the field facts ; their bearing upon the problem of 
rock origin seems worthy of consideration. 

Of the two contrasting types of rock — the calcic and 
the alkalic — the former appear to have been developed 
by the simpler processes. The quartz-rich calcic con- 
tact phases of the Durham quartz diorite seem quite 
clearly the result of the assimilation of Kittery quartz- 
ite by the gabbroid magma. The study of a number 
of selected specimens in which fragments of a quartzite 
were seen in various stages of alteration seems to prove 
this point. The abundance of apatite, indicating the 
presence of plenty of mineralizer, would, moreover, 
appear to show that the assimilation was aided by the 
action of the rock's volatile components. But the grada- 
tion from a gabbroid margin to a granitic center seems 
hardly the result of assimilation as no granite was seen 
in the marginal phase. The formation of granite might 
in part result from a process of fractional crystallization 
under gravitative control as the laboratory work of 
Bowen 18 indicates may take place experimentally. But 
such a simple setting is hardly the entire story. In 
studying by means of thin sections several suites of 
specimens taken across the Durham body, two facts 
stand out rather clearly: (1) the gradual disappearance 
of pyroxene and the development of hornblende and 
biotite, both minerals that all experimental work indi- 
cates can only be formed in the wet way; (2) the partial 
disappearance of calcic feldspar and the development of 
sodic and potassic feldspar, the latter again, judging 

18 Jour. Geology, vol. 23, No. 8, suppl., 1915. 

Portsmouth Basin, Me. and N. H. lot 

from experimental evidence, only to be produced in the 
wet way. Tims to the theory of fractional crystalli- 
zation under gravitative control a necessary addition 
would seem to be the influence of mineralizing solu- 
tions in furthering changes in composition. It is entirely 
probable that for a given locality either of these agencies 
may have been chiefly responsible for a given rock type. 
In the case of the Durham body, since it shows marked 
contact effects, it would seem necessary to assign an 
important role to the action of the volatile components 
in producing the granite. 

The alkaline rocks offer a more complex problem. In 
this case a biotite granite is succeeded by alkaline granite 
which in turn is apparently followed by alkaline rocks 
carrying less and less quartz but rich (10-13%) in such 
elements as soda and potash. Here and there in these 
alkaline rocks are found coarse pegmatitic phases in 
which the crystals of feldspar measure up to 10 centi- 
meters in length and the alkaline hornblende 5 to 6 centi- 
meters. The development of soda — and potash — rich 
minerals and the presence of pegmatitic phases are both 
interpreted as indicating that mineralizers were active 
agents in producing these peculiar rock types. There 
are, however, other factors, which must be reckoned with. 
The experimental work of Bowen indicates that frac- 
tional crystallization under gravitative control may 
largely influence the process of differentiation. The 
contacts of the subjacent bodies against the sedimen- 
taries certainly indicate that magmatic stoping was one 
of the means of emplacement. The figures dealing with 
specific gravities appear to show that much of the stoped 
material would sink. With stoping a factor of emplace- 
ment, then the assimilation of the stoped blocks would 
seem necessary since the contact actions show that the 
process actually takes place. Such stoped and assimi- 
lated material ought to make itself felt as the process of 
differentiation continues. Both of these aforementioned 
processes have evidently exerted a control upon the 
differentiates, but to the writer it seems that their effects 
were furthered by volatile components which not only 
aided assimilation but also were active in producing the 
upward transfer of the alkaline constituents so necessary 
in producing the alkaline rock types. 


A. Wandke — Intrusive Rocks. 

Correlation of the Intrusives. 

Alkaline rocks were at one time regarded as rare in 
New England, but little by little new localities for the 
occurrence of these types are being reported. The 
subjoined table notes some of the localities in New 
England from which alkaline rocks have been reported 
and also shows the approximate ages of the intrusives. 


Mt. Ascutney, 

Cuttingsville, 20 

Litchfield, 21 

Belknap Mountain, 22 
New Hampshire. 

Types of Intrusives. Age. 

Biotite granite, nord- Post Carboniferous and 
markite, pulaskite, pre-Cretaceous. 
diorite, essexite, gab- 
bro, paisanite. 

Essexite and nepheline Age unknown, 

Syenites and nepheline Age unknown, 

Syenite and aplite. 

Tripyramid Mountain, 23 Monsonite and syenite. 
New Hampshire. 

Blue Hills, 24 

Copper Mine Hill, 2 
Ehode Island. 

Essex County, 26 

Portsmouth Basin, 
Maine and New 

No age given. 
No age given. 

Upper Carboniferous. 
Middle Devonian. 
Middle Pennsylvania!!. 


Quartz diorite and bio- 
tite granite. 


Diorite, calci-alkaline 

granite, grano-dior- 

ite, gabbro diorite. 
Alkaline syenites, Cape 

Ann and Peabody 

granites, Andover 

alkaline granite. 

Biotite granite. 

Quartz diorite, biotite 
granite, gabbro, 
grano-diorite, alka- 
line granite, alkaline 
syenite, gabbro? 

Foxgrove House, Cambridge, Mass. 

19 E. A. Daly, U. S. Geol. Survey, Bull. 209, 1903. 

20 J. W. Eggleston, this Journal, vol. 45. 1918. 

21 E. A. Daly, Bull. Geol. Soc. Amer., vol. 29, p. 463, 1918. 

22 L. V. Pirsson, this Journal, vol. 22, 1906. 

23 L. V. Pirsson and W. N. Eice, this Journal, vol. 31, 1911. 

24 C. H. Warren and S. Powers, Bull. Geol. Soc. of Amer., vol. 25, p. 463. 
23 C. H. Warren and S. Powers, ibid. pp. 452 and 463. 

20 C. H. Clapp, U. S. Geol. Survey, Bull. 704, 1921. 

Post-Ordovician and pre- 
Silurian ? 

Mississippian or Pennsyl- 

Post Upper Carbonifer- 

W at anabe— Babingtonite from Japan. 159 

Art. XV. — On the Babingtonite from the Contact Meta- 
morphic Deposits of the YakuJci Mine, Province Iivaki, 
Japan; by Manjiro Watanabe. 

Mode of occurrence. 

The Yakuki mine is located near the eastern margin of 
the Abukuma mountainland, 1 in the Province of Iwaki, in 
northeastern Japan. It has been worked for chalcopyr- 
ite in the skarn 2 masses, which were formed in the contact 
aureole between granodiorite and the Paleozoic clay slate 
and limestone. The minerals, constituting the ore 
deposits, are chiefly hedenbergite and garnet with some 
chalcopyrite. Magnetite and sphalerite are also found 
in notable amounts in some portions. The babingtonite, 
to be described, occurs in veins and irregular aggregates, 
mainly in the hedenbergite masses in the western part of 
the mine. In some specimens, these masses of babington- 
ite, cut by garnet, are brecciated into small fragments. 
Some of these fragments are surrounded by thin layers 
of babingtonite, on which minute crystals of garnet are 
planted. Thus, it is clearly seen that these two minerals 
were repeatedly deposited one after the other. 

These fragments in the brecciated zone are usually 
cemented by quartz or calcite. In the latter case, the 
crystals of babingtonite, planted on the fragments, are 
easily isolated by dissolving the carbonate in acid. 

In other cases, the babingtonite appears in minute vein- 
lets, which cut the hedenbergite, magnetite and ilvaite. 3 
Such veinlets often thin out in one direction. The wider 
portion is filled with garnet, and the narrower portion is 
made of garnet and babingtonite, the former mineral 
appearing implanted on the walls. In still narrower 
parts, babingtonite occurs with quartz, and the veinlet 
terminates as a pure aggregate of quartz. 

1 B. Koto: The Archean formation of the Abukuma Plateau, Jour. Coll. 
Sci. Imp. Univ. Tokyo, vol. 5, Pt. 3, 1893. 

2 This term was originally a Scandinavian miner 's word, which means rub- 
bish. However, it was applied by V. M. Goldschmidt to a certain type of 
contact metamorphic rock, which consists essentially of coarse-grained aggre- 
gates of pyroxene and garnet, and is now widely used with this meaning. 

3 This mineral is found in very small amount, and is not accurately deter- 
mined. It is black in color, and is generally opaque, but transmits light in 
very thin portions, where the pleochroism is marked. It is easily attacked 
by hydrochloric acid. It is most probably ilvaite. 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 20.— August, 1922. 

160 Watanabe — Babingtonite from Japan. 

When the babingtonite vein traverses magnetite or 
ilvaite, it assumes a shape of a regular vein, but as it cuts 
hedenbergite, babingtonite often replaces a part of heden- 
bergite on both sides of the fissure, and still preserves the 
original texture. 

In short, the formation of the babingtonite in the 
Yakuki mine seems to have commenced at a certain stage 
of the formation of skarn masses, and to have continued 
after the brecciation of the skarn and lasted to some 
stages of fissure, filling deposition of quartz and calcite. 
Throughout the whole duration, the mineral is most inti- 
mately associated with garnet. 


The isolated crystals were picked out from the speci- 
mens, in which the mineral is embedded in calcite, and 
were measured by means of a Czapski two-circle goni- 
ometer. 4 Most of these crystals are very small, rarely 
exceeding 2 or 3 millimeters in dimension. Besides, the 

Table I. 


Calculated 5 







c 001 

124 c 








o 010 







a 100 









h 110 


















k 110 









o 011 









s 011 









d 101 









n 101 









p 111 









; 102 









I 201 









q 111 









z 023 









4 This study was done through the courtesy of Professor A. F. Eogers 
of Stanford University, Calif., to whom the writer offers his sincere thanks. 

5 The calculations of the co-ordinate angles were made mainly according 
to the formulae given by A. J. Moses and A. F. Eogers (School of Mines 
Quarterly, vol. 24, No. 1, 1902), and partly by the method, proposed by H. E. 
Boeke (Die Anwendung der stereographischen Projektion bei kristallogra- 
phischen Untersuchungen, 1911, pp. 41-46). 

Watanabe — Babingtonite from Japan. 161 

crystal faces are often striated and curved, and the 
appearance of double and multiple images makes it 
doubtful which is to be measured. In spite of such diffi- 
culties in measuring, the mean values obtained by 
measurement closely agree with the calculated ones. 

The observed forms and their co-ordinate angles are 
shown in Table I. 

The four forms marked with an asterisk are new to this 
mineral. The crystal orientation and elements, used for- 
notation and calculation, are those of Dauber, 6 the advan- 
tage of which has been stated by Palache and Fraprie. 7 
It is necessary, therefore, to change the symbols of forms, 
when these results are compared with the descriptions, 
which are found in the treatises of Dana and Hintze, or 
with those of Goldschmidt, 

The calculated and measured angles of the forms in 
the vertical zone agree closely; but with regard to the 
terminal faces, the deviation of the measured values from 
the calculated ones is often great. This is partly due to 
the general disadvantage of a theodolite goniometer that 
the possible error by measurement of the meridional 
angle 4> increases as the polar distance p diminishes. To 
compensate this deficiency, the interfacial angles between 
the faces belonging to the zone (be) were measured by 
using the theodolite goniometer as a one-circle one. The 
results of these measurements fairly well coincide with 
the calculated values as shown in Table II. 

Table II. 

Calculated. (mean). 

be 92° 36' 92° 29' 

be' 87° 24 87 32 

bs 44 40 44 08 

cs ..-'.... 42 44 43 07 

co 45 13 45 22 

bo 47 23 47 25 

6 The crystal elements of babingtonite, given by Dauber (Pogg. Ann., 94, 
402, 1855) are as follows: 

a : b : c = 1.635 : 1 r 0.8955; a = 86° 09', p = 93° 48', y — 112° 22'. 
ab — 67° 48', ac = 87° 28', be — 92° 36'. 

These values were obtained by him from the results of measurements on 
more than eighty crystals of babingtonite- from Arendal. 

7 Palache and Fraprie, Proc. Amer. Acad., 38, 382, 1902. Some misprints 
occur in their table of corresponding forms for the different positions. 

162 Watanabe — Bdbingtonite from Japan., 

The crystal habits and the zonal relations of the faces 
are illustrated in the accompanying figures. 8 

Among the observed forms, both b and k are generally 
well developed and give a prismatic habit to most crystals, 
although some platy crystals are produced by the pre- 
dominance of b. These two forms are usually character- 
ized by vertical striations, caused by their mutual oscil- 
latory combination. The forms a, h and g all appear as 
narrow faces, which truncate the acute edges between b 
and k. Among the terminal faces, c is best developed in 
size, though it is striated parallel to the axia a, owing to 
the oscillatory combination of c and d. All other forms 
are represented as either narrow or minute faces. 

Physical Properties. 

The mineral is black in color and has a brilliant vitre- 
ous luster. Crystals are commonly opaque, but in thin 
sections transmit light and gives deep color. The pleo- 
chroism is very marked, the axial colors being as follows : 

X deep emerald-green, 

Y purple-brown, 

Z , deep brown. 

In this particular point, the mineral is distinguished 
from other triclinic members of the pyroxene group, 
which have the similar crystal forms. The absorption 
is strong in the order of X > Y > Z, and the dispersion 
is marked. Hence, it is difficult to accurately determine 
the extinction position. As the averages of repeated 
measurements on several sections, the extinction angle 
on 6(010) is given as about 37° from the vertical axis c, 
and that on c(001) is about -5° or -6° from the axis a. 

8 In drawing these figures, a gnomonic projection of the crystal on the 
plane normal to the prism zone was first made (fig. 1). Each edge in the 
plans on this plane (figs. 2b and 3b) is normal to the zone line, which con- 
nects the poles of the faces intersecting at that edge. Next, to make the 
elevations on the plane of fc(010) (figs. 2c. and 3c), the intersection of this 
plane with the plane of projection and its angle point were found on the 
projection. In this particular case, they are shown as the line LL' and the 
point B respectively. Each edge in the elevations is normal to the line con- 
necting this point B and the intersection between the line LL' and the zone 
line, which passes the poles of the two faces intersecting at that edge. Like- 
wise, the line LL' and its other angle, point B', determine the directions of 
the edges of elevations on the plane of b' (010) (figs. 2a and 3a). (See 
Boeke: Die gnomonische Projektion, 1913, pp. 40-44.) 








- 1 — _ 


1 >^ 



• i / 

i / 

/* r'lv. 


/>\ jiV 

V 3J 

/ ,/ _^><Cj°A 

\\ BIX^ 



l<t oo\ 




-^^x /°' u 

\ ^^>\ / 



y/ P\\ 




' 1 "\ 

1 \ 



V a N 


fig. 3a. 

Fig. 1. — Gnomonie projection of the babingtonite from the Yakuki mine, 

Figs. 2 and 3. — The orthographic drawings of the babingtonite from the 
Yakuki mine. 

Figs. 2a and 3a. — Elevations on the plane of (010). 

Figs. 2b and 3b. — Plans on the plane vertical to the prism zone. 

Figs. 2c and 3c. — Elevations on the plane of (010). 

164 Watanabe — Babingtonite from Japan. 

The indices of refraction were measured by Becke's 
method by means of the standard solutions, which were 
available through the courtesy of Prof. A. F. Rogers. 
The minimum value is between 1.710 and 1.720, and the 
maximum a little greater than 1.740. The intermediate 
principal index lies between 1.720 and 1.730. 9 Thus the 
three principal indices are approximately: 

a = 1.715±0.005 
fi = 1.725±0.005 
7 = 1.740+O.OOx 

Optically biaxial and positive. Cleavage is perfect 
parallel to c(001), and less perfect parallel to a (100) and 
&(010). The hardness is almost equal to that of ortho- 
clase, or about 6. Before the blowpipe, the thin edges 
easily fuse to magnetic globules. The substance is not 
affected by common acids. 

Tohoku Imperial University, Sendai, Japan. 

9 Determined by S. Tsuboi's method, Jour. Geol. Soc. Tokyo, vol. 25, p. 38, 

0. Holtedahl—A TUlite-lilce Conglomerate. 165 

Art. XVI. — A Tillite-like Conglomerate in the "Eocam- 
brian" Sparagmite of Southern Norway; by Olaf 


The so-called Sparagmite formation, which covers a 
very large area of southeastern Norway (see fig. 1), is in 
several respects a sedimentary series of considerable 
interest. TVe meet here with thick beds of coarse-grained 
clastic rocks, very rich in fresh feldspar, alternating with 
thinner zones of clayey material and limestones. Of 
these the Biri limestone is of most importance and best 

Xo fossils have as yet been found in these rocks and 
therefore there are somewhat divergent opinions as to 
their exact age. We know that the sparagmites are older 
than the Lower Cambrian Holmia shale, and as there 
is a good transition from the highest sandstone into 
the Holmia shale, Brogger, Minister, and more recently 
J. Kiaer have regarded the sparagmites as closely attached 
to the Cambrian. Brogger introduced for them the term 
"Eokambrruiii," thus indicating that the strata are of the 
oldest Cambrian, while Kiser classifies them as true Lower 
Cambrian. The Swedish geologist Tornebohm, on the 
contrary, referred the sparagmites to the Algonkian. 

Xon- Scandinavian authors have also discussed the age 
of the Sparagmite series or parts of it. Walther, in his 
paper "LVber algonkische Sedimente," 1 has emphasized 
the great petrological likeness to the Torridonian of Scot- 
land, while Rothpletz, 2 by assuming an overthrust that 
quite certainly does not exist, held that the Biri limestone 
is younger instead of older than the Holmia shale. From 
this viewpoint of Eothpletz, Grabau, in his "Comparison 
of American and European Lower Ordovicic Forma- 
tions," 3 has discussed the possibility of the Biri limestone 
being a continuation of the Durness limestone of north- 
ern Scotland. 

The main objects of the present article are: (1) to 
point out the occurrence in the sparagmites of conglom- 
erates of a tillite-like character, and (2) to give a sum- 

1 J. Walthier, Zs. deutseh. Geol. Ges., 61, 2S3-305, 1909. 

: A. Eothpletz, Sitzber. Bavr. Akad. Wiss., 1-66, 1910. 

3 A. W. Grabau, Bull. Geol.' Soc. America, 27, 555-622, 1916 

166 0. Holtedahl — A Tillite-like Conglomerate. 

mary review of the stratigraphic sequence of the series, 
from which the chronologic position of the conglomer- 
ates may be seen. A great many Scandinavian geologists 
have contributed to the clearing up of these features, as 
Kjerulf, Schiotz, Tornebohm, Minister, Bjorlykke, Gold- 
schmidt, Werenskiold, and others. The writer has 
studied the problem of the Sparagmites in the field dur- 
ing the summers of 1919 and 1920. 

Fig. 1. — Sparagmite area of Southern Norway dotted. M=zMjosen. 

The greater and especially the northern and central 
parts of the region covered by sparagmite rocks have been 
very decidedly deformed by the Caledonian orogeny, and 
the rocks are highly metamorphosed. The stratigraphic 
sequence given below is based especially on the conditions 
met with in the most southern belt, particularly in the 
region around the northern part of the great lake Mjosen. 
Even here, intense folding and thrusts have taken place, 

0. HoltedaliJ — A Tillite-like Conglomerate. 167 

making the deciphering of the geologic history a rather 
complicated matter. 

The oldest members of the Sparagmite division are 
known from the more northern district only. Here, as for 
instance between the Gudbrandsdal and Oesterdal, abont 
75 km. north of Mjosen, the oldest sparagmite is a some- 
what metamorphosed rock, having a conglomerate at the 
very base, and is seen to rest on the somewhat nndnlating 
surface of pre-Cambrian gneiss. In the southern part of 
the area the oldest zone, the gray or older sparagmite, 
several hundred meters in thickness, is a dark gray, gen- 
erally coarse sandstone rich in grains of feldspar, with 
some layers of dark arenaceous shale. Then follows a 
relatively thin zone with red and greenish shale and thin 
beds of limestone. The earlier time of strong denudation 
and rapid sedimentation of the detritals derived from 
granites and gneisses, of which the thick sparagmite zone 
tells, changed later into one with only slow deposition in a 
playa-like basin of water. 

Denudation again became active, indeed to a quite 
remarkable extent, for above the last-mentioned zone 
there is a very coarse conglomerate, in places 100-200 
meters thick, the Biri conglomerate, consisting of bowl- 
ders very often of large size, up to 1 meter in length. 
The bowlders, made up of granite, gneiss, quartzite, dia- 
base, and limestone, are well rounded and distinctly 
water-worn. A river transport of such coarse material 
certainly presupposes a relatively steep grade of the land. 

Through an intermediate zone of gray sparagmite and 
shale, the sequence passes into the Biri limestone, the 
thickness of which in places is as much as 100-150 meters. 
This zone in part consists of a compact gray (generally 
more or less magnesian), often arenaceous limestone, the 
main constituent being, however, a black argillaceous 
limestone. In the compact variety there are at Lake 
Mjosen intraformational conglomerates, with angular 
pieces of limestone lying at very different angles, thus 
indicating very shallow waters. There are also oolitic 
and at one locality stromatolitic limestones, which appear 
to have been chemically precipitated. In spite of much 
seeking for fossils, none were found. 

Over the limestones then follows an accumulation of 
huge masses of coarse elastics, the younger, or red, spar- 

168 0. Holtedahl — A Tillite-like Conglomerate. 

agmite, a pink, coarse, very often conglomeratic, thick- 
bedded sparagmite, 300-400 meters thick, and in places 
nrnch more. As to the conditions nnder which this and 
the other beds of sparagmites have been formed, the large 
amount of fresh feldspar necessarily presupposes, as has 
been pointed ont by Walther and by several Norwegian 
geologists, a weathering of granitic rock in a very dry 
climate. That we are not dealing with wind-blown desert 
sands is evident from the coarse and angular character of 
the mineral grains, as well as from the absence of dune 
stratification. There was here a very short and very 
rapid detrital transport by torrential waters, from encir- 
cling mountains into an intermontane basin, under semi- 
arid conditions. The thickness of the beds, e. g., the red 
sparagmite, the coarseness of the material, and the high 
content of feldspar grains seemingly indicate a far shorter 
length of transport and a much more rapid accumulation 
than do the ordinary red sandstones of later formations. 

The red sparagmite is often seen to have conglomerate 
zones with well rounded, water-worn bowlders of moder- 
ate size. This passes upward into a more peculiar con-^ 
glomerate to which I should like to draw special atten- 
tion. It has a dark reddish brown color, with no visible 
bedding at all and with fragments of all sizes embedded 
in a fine-grained matrix. The smaller fragments are 
generally quite angular and often the larger bowlders, 
some of which are more than 0.5 meter in diameter, are 
only partially rounded. The greater part of the frag- 
ments consist of granites, gneisses, sandstones, and 
quartzites. In addition, there are limestones, as in the 
western district where this conglomerate is found at Lake 
Mjosen, while in an eastern area, north of Trysil near the 
Swedish frontier, occur porphyries of types known from 
the pre-Cambrian still farther east. The limestone 
bowlders mentioned are quite like those previously spoken 
of as occurring in the Sparagmite division. 

The conglomerate just described 4 has all of the general 
characteristics of a tillite, and it is difficult to think of any 
other method of transportation for material of this kind 
than a glacial one. As yet, however, no undoubted glacial 

4 An illustration of a specimen of the conglomerate is given in a paper by 
the present author, ' ' Om Trysilsandstenen og sparagmif avdelingen, ' ' Norsk. 
Geol. Tidsskrift, 6, 38, 1920. 

0. Holtedalil—A TiUite-like Conglomerate. 169 

striae have been observed, but since the sparagmites have 
suffered much from the Caledonian deformation, the find- 
ing of striae is very difficult. In fact, when the bowlders 
are freed from the matrix, their surface is generally seen 
to be slickensided and tectonically striated. This con- 
glomerate is exactly like the brown tillites of Finmarken, 
recently mentioned by me in this Journal. 5 

As there is no break in the sequence below the conglom- 
erate just described, it can not have been deposited by a 
glacier on the dry land ; and if we assume a glacial origin, 
the material must be thought of as having been deposited 
by floating ice, or, better, as morainic material in front of 
a submerged glacier. 

Above the tillite-like conglomerate, the thickness of 
which at Lake Mjosen appears to be at least 10 meters 
(perhaps considerably thicker), though somewhat less in 
the eastern district, follows a relatively thin zone of red 
and green shale, sometimes with sandstone beds, and 
this in turn is overlaid by the so-called "Quartz sand- 
stone." As the name indicates, this formation is a nor- 
mal sandstone, consisting essentially of quartz with the 
mineral grains smaller, more even-sized and rounded than 
in the sparagmites ; zones of sparagmite-like rock, how- 
ever, are also met with here. The color of the quartz 
sandstone is mostly from bluish to yellowish gray, and 
the material indicates a longer transport, with better sort- 
ing and rounding of the grains. In the southwestern part 
of the sparagmite area this sandstone is very thick, cer- 
tainly upward of 100 meters. It gradually passes into 
sandy shale, which in turn passes into greenish shale 
holding the Holmia fauna which has recently been revised 
and described by Kiaer. 6 The stratigraphic succession 
is illustrated in generalized form in fig. 2. 

If we next consider the distribution of the different 
formations above described, it becomes evident, as has 
been made especially clear by the author's recent studies 
along the eastern border of the sparagmite area, that the 
quartz sandstone in the southern part of the area has a 
much wider distribution than the older feldspar-bearing 
zones. Ten kilometers west of the Engerdalen valley the 

5 0. Holtedahl, this Journal (4), 47, 104, 1919. 

c J. Kiaer, The Lower Cambrian Holmia Fauna at Tomten in Norway, 
Vidensk. selsk. Skr., 1916, Xo. 10. 

170 0. Holtedahl — A Tillite-like Conglomerate. 

sequence is in the main as described above, for here a 
thick mass of red sparagmite is exposed as the oldest 
zone, while in the valley itself the lower part of the 
"quartz sandstone" is developed as a true, and in part 
a very coarse, sparagmite (the "gray sparagmite" of 
0. E. Schiotz), resting directly on pre-Cambrian granite. 
Here denudation must have taken place while the red spar- 
agmite was being deposited nearby to the west. On the 
other hand, the red sparagmite of the eastern localities 
mentioned, that is, in an area that must be very close to the 
eastern boundary line of this formation, is not in any 
respect different from that farther west. In other words, 
it does not seem likely that the present boundary line of 

Fig. 2. 
S. W. N. E. 

Fig. 2. — Generalized stratigraphic succession of the southern sparagmite 
area and adjacent districts before the Caledonian deformation (Ordovician 
and Silurian divisions not introduced). Complete sequence in middle part 
of the section: 

Upper and Middle Cambrian (alum shale). 

Lower Cambrian Holmia shale (in black). 

Quartz sandstone. 

Eed sparagmite with tillite-like conglomerate on top. 

Biri limestone. 

Gray sparagmite. 
To the left, pre-Cambrian gneiss and granite. To the right, pre-Cambrian 
granite, porphyry, and Trysil sandstone with sheets of diabase. 

the red sparagmite was the original one, because then we 
should expect distinctly coarser rocks at the eastern local- 
ities which were nearer to the highlands. Moreover, along 
the western boundary of the sparagmites the red sparag- 
mite (and its metamorphic equivalents) are seen suddenly 
to disappear in this direction. If we remember, further- 
more, that in the Mjosen district, near the southern 
boundary of the sparagmite area, bowlders of limestone, 
which no doubt belong to the lower part of the sparag- 
mites, occur in the tillite-like conglomerate at the top of 
red sparagmite, we must conclude that there have been 
decided marginal uplifts in younger Sparagmite time, 

0. Eoltedahl — A Tillite-like Conglomerate. 171 

causing a peripheral denudation through the sparagmites 
down to the granite and porphyry. With an especially 
marked marginal uplift, glaciers and a laying down of 
conglomerates without stratification may have come into 
existence. It seems natural to assume that this uplift 
represents not merely a gentle flexure, but a decided 

When the deposition of the quartz sandstone began, the 
land to the east of the red sparagmite was well base-lev- 
elled, while toward the southwest there were still rising 
heights. And as was shown in the paleogeography map 
of Lower Cambrian time published by the writer in 1920 7 
there was still land in this direction at the time of the 
Holmia shale. 

That marginal upheavals or central sinkings have 
repeatedly taken place in the area here considered, and 
in the time of the sparagmites, preceding the deposition 
of the oldest known fossiliferous sediments, is evident 
from the general geologic occurrences. Such thick 
accumulations of coarse materials, with their distinct 
petrological characters, can only have been transported 
over very short distances. That this upheaval has been 
not a gentle warping but, rather, an important dislocation 
of the crust is, in my opinion, further indicated by the 
large bowlders found above the older sparagmite. A con- 
glomerate like this one, following the deposition of spar- 
agmitic sandstones and arenaceous shales, can scarcely 
have been brought about alone through climatic changes ; 
it must be due to a decided uplift of the adjacent land. 

Even though there must have been quite remarkable 
vertical crust movements during Sparagmite time, we 
know of no folding causing distinct unconformities. 
On the other hand, a distinctly folded sedimentary series, 
not far from the sparagmites, is found in the Trysil sand- 
stone, which is the direct western continuation of the 
Dala sandstone of the Swedish geologists and a repre- 
sentative of the Jotnian formation of the pre-Cambrian 
of Fennoscandia. The Trysil-Dala sandstones and the 
Sparagmite have previously been generally considered to 
belong to one great stratigraphic group, but to me this 
appears to be a wrong correlation; the former clearly 

T 0. Holtedahl, this Journal (4), 49, 3. 

172 0. Holtedahl — A Tillite-like Conglomerate. 

belong to the pre-Cambrian, while the latter are closely 
attached to the Holmia shale of the Lower Cambrian. 
The folding of the Jotnian sandstone, however, has not 
been decided; indeed, over the Swedish frontier it is 
sometimes almost without folding. On the other hand, 
considerable crust movements did occur in Sparagmite 
time. In fact, we appear to have here the final dying out 
of the pre-Cambrian diastrophic movements, the more 
local vertical movements following the regional and much 
older tangential ones. In some ways it may be said that 
the sparagmite s are of a time transitional between the 
true Proterozoic and the Paleozoic, even if stratigraphi- 
cally they are best classed with the Paleozoic. 

It is only natural that sediments like the sparagmites 
should give rise' to divergent opinions as to their age 
relations. In similar fashion, the well-known tillites of 
China and Australia, which occur below fossiliferous 
Cambrian, have been classified both as Cambrian and pre- 
Cambrian. Such a struggle is often a struggle of words 
only. The main thing is that we should have a sound 
idea as to the relative stratigraphic position of these 
ancient tillites and of their relation to the times of 
orogeny. Their age relations will then in the course of 
geologic endeavor be satisfactorily determined. That 
the tillite-like conglomerate described in this paper is 
of an age similar to that of the eastern tillites just 
mentioned seems at any rate to be a rather probable 

Table of pre-Ordovician formations in the southern sparagmite 
area of southern Norway. 

(Also compare with. Fig. 2.) 
Paradoxides shale. Mid- Invasion of sea in south and west of 

die Cambrian. sparagmite area, 

Holmia shale. Lower Final pre-Ordovician denudation and 
Cambrian. base-levelling of land south and west 

of sparagmite area. Invasion of sea 

into sparagmite region and further 


Quartz sandstone. Very strong denudation of land south- 

100 + m. west of sparagmite area ; thinning out 

of sandstones toward northeast. 
Upper red and green Slow denudation and final peneplaria- 
shale. Thin Eocam- tion of land east of area of red sparag- 
brian. mite. 

0. Holtedahl—A TiUite-Uke Conglomerate. 173 

Tillite-like conglomerate. Marginal upheaval (or central sink- 
At least 10 m. ing) and very strong denudation of 

surrounding land, possibly by gla- 
Red, coarse, y o u n g e r Strong weathering and denudation of 
sparagmite, often con- marginal lands, periodic flood trans- 
glomeratic. 300-400 m. port of detrital material into central 

Biri limestone. Quiet conditions; deposition of cal- 

About 100 m. careous material. 

Biri coarse c o n g 1 o m- Exceedingly strong denudation of 
erate. 100-200 m. surrounding land, probably caused by 

important marginal uplift. 
Lower red and green Slow sedimentation of argillaceous 
shale, with beds of thin and calcareous material, 
Gray, older sparagmite. Strong weathering and denudation of 
Several hundred me- marginal lands, periodic flood trans- 
ters. port of detrital material into central 


Older sparagmite rocks known in more northern part of spar- 
agmite area. 

Total maximum of Eocambrian sparagmites more than 2000 m. 
Probably a lacustrine deposit on a low land bordering the sea. 

Erosion interval and unconformity. 

Intrusion of diabases into, and folding of, Jotnian Trysil sand- 

Trysil sandstone. 

Porphyry, granites, gneisses. 


Dr. Alfred Goldsborough Mayor, for many years a Director 
of the Carnegie Institution of Washington, died at the Tortugas, 
Key "West, Florida, June 24, at the age of fifty-four years. He 
was the son of Professor Alfred M. Mayer, the eminent physicist 
of Lehigh University and later of the Stevens Institute of 
Technology ; the results of many of Professor Mayer 's researches, 
particularly in acoustics, have been published in this Journal. 

Dr. Mayor's early training was in the field of mechanical engi- 
neering and, at the conclusion of his studies in this subject, he 
received the degree of M.E. from Stevens Institute in 1889. 
After graduation he became assistant in physics first at Clark 
University and then at the University of Kansas. But after 
three years of this work, he found his interest in the mathemati- 
cal sciences was overshadowed by his intense love of natural his- 
tory, and he therefore returned to the east and became associated 

174: Obituary. 

with Alexander Agassiz at Harvard. He served as Mr. Agassiz 's 
assistant and was given charge of the collection of radiates in the 
Museum of Comparative Zoology. He retained this position for 
eight years, accompanying Mr. Agassiz on several expeditions to 
the tropical seas and acquiring that knowledge of marine life and 
the methods for its investigation which fitted him so well for the 
special line of work which occupied his more mature years and 
in which his reputation as a naturalist has been made. While at 
Harvard he completed the requirements for the degree of Sc.D., 
the degree being conferred in 1897. 

In 1900 Dr. Mayor became curator-in-chief of the museum of 
the Brooklyn Institute as well as curator of natural sciences. 
Four years later he was appointed director of the department of 
marine zoology of the Carnegie Institution, in which position 
he has contributed his most important services to biology. He 
has also served as lecturer in biology at Princeton University 
since 1913. 

The region where the Gulf Stream impinges on the southern 
coast of Florida has long been known as one of the richest regions 
of the world in the variety and abundance of its marine life. 
In 1902, under the auspices of the Brooklyn Institute, Mayor 
led an expedition to this locality, with which he was already 
familiar, and hit upon the brilliant idea of utilizing the govern- 
ment reservations of the Tortugas islands as a marine biological 
laboratory. This plan was eventually supported by the Carnegie 
Institution and through the enthusiasm and self-sacrificing 
leadership of Dr. Mayor as director matured into the permanent 
marine station which has ever since been available to the zoolo- 
gists of the world. At this laboratory many of the most prom- 
inent zoologists have gained their first acquaintance with 
tropical marine life and their studies here inaugurated have 
yielded results of far-reaching importance. 

Mayor's direct contributions to Zoology consist of numerous 
papers mainly on coelenterates, which he studied intensively 
both anatomically and physiologically, His most extensive pub- 
lication, in three quarto volumes, is a superbly illustrated mono- 
graph on the medusas of the world. To secure material for this 
work he accompanied Mr. Alexander Agassiz for many years in 
explorations of the tropical seas of both the Atlantic and Pacific 
Oceans. He was later the leader of expeditions to the coral 
reefs of Samoa, Fiji and other islands of the tropical Pacific. 

Mayor has also made contributions of importance on the evolu- 
tion of snails, the coloration of insects, the nature of the nervous 
impulse, the formation of coral reefs, the rate of growth of 
corals, the palolo worm, and studies on the physical nature of 
death. He had a brilliant mind, a friendly, modest and unselfish 
personality, and a restless energy in his work; he naturally 
became a member of many learned societies and an officer of 
several. w. R. c. 

Ward's Natural Science Establishment 

A Supply-House for Scientific Material. 

Founded 1862. Incorporated 1890. 

A few of our recent circulars in the various 
departments : 

Geology : J-32. Descriptive Catalogue of a Petrographic Col- 
lection of American Rocks. J- 188 and supplement. 
Price-List of Rocks. 

Mineralogy: J-220. Collections. J-238. Minerals by Weight. 
J-224. Autumnal Announcements. 

Paleontology: J-201. Evolution of the Horse. J-199. Palae- 
ozoic index fossils. J-115. Collections of Fossils. 

Entomology : J-33. Supplies. J- 229. Life Histories. J-230. 
Live Pupae. 

Zoology: J-223. Material for dissection. J-207. Dissections 
of Typical Animals, etc. J-38. Models. 

Microscope Slides: J-189. Slides of Parasites. J-29. Cata- 
logue of Slides. 

Taxidermy: J-22. North American Birdskins. Z-31. General 

Human Anatomy: J-37. Skeletons & Models. 

General : J-228. List of Circulars & Catalogues. 

Ward's Natural Science Establishment 

84-102 College Ave., Rochester, W. Y., V. 8. A. 


number consisting of 100 to 120 pages). Editor : EUOENIO RIGNANO. 

This is the only review which has a really international collaboration ; which is 
of world-wide circulation and occupies itself with the synthesis and unification of 
knowledge, in the history of the sciences, mathematics, astronomy, geology, physics, 
chemistry, biology, psychology and sociology. 

It is the only review which, by inquiries among eminent scientists and writers 
(on: The philosophical principles of the sciences; Fundamental astronomical and physical 
questions of current interest; Contributions given by the various countries to different 
branches of knowledge; Question of vitalism; Social question; International questions 
raised by the world war), makes a study of the problems interesting scientific and intel- 
lectual circles throughout the world. 

It has published articles by Messrs. : Abbot = Arrhenius -Ashley - Bayiiss- Beichman- 
Bigourdan - Bohlin - Bohn - Bonnesen - Borel - Bouty - Bragg - Bruni - Burdlck - Carver - 
Caullery - Chamberlin - Charlier = Claparede - Clark - Costantin - Crommelln - Crowter - 
Darwin •> Delage - De Vries - Durkheim = Bddington - Edgeworth - Emery = Enriques - 
Fabry - Pindlay - Fisher - Fowler - Golgi - Gregory - Harper - Hartog - Heiberg - Hinks - 
Hopkins-Inigues-Innes-Janet-Kaptein-Kaye-Kidd-Langevin-Lebedew-Lloyd Morgan - 
Lodge - Loisy - Lorentz - Loria -Lowell - MacBride - Melllet =Moret-Muir=Peano -Picard - 
Poincare-Pulseux-Rabaud-Rey Pastor- Righi-Rignano-Russell-Rutherford-Sagnac- 
Sarton- Schiaparelli - Scott - See - Sherrington - Soddy - Starling- Svedberg- Thomson - 
Thorndike -Turner -Volterra -Webb -Weiss -Zeeman and more than a hundred others. 

"SC1ENTIA" publishes its articles in the language of its authors, and joins to the 
principal text a supplement containing the French translations of all the articles that 
are not in French. (Write for a Specimen Number to the General Secretary of 
"Scientia", Milan, sending, -to defray postal and other expenses,- 2 Francs in 
stamps of your country.) Annual subscription : 40 sh., or 10 dollars post free. 

Office : 43 Foro Bonaparte, Milan, Italy 

Publishers: W.LIAMS & NORG ATE - London ; FELIX ALCAN- Paris 


Art. VIII. — Colloids in Geologic Problems ; by G-. D. Hubbard, 95 

Art. IX.— Primitive Pecora in the Yale Museum; by R. S. 

Lull, Ill 

Art. X. — A Critical Phase in the History of Ammonites; by 

C. Diener, 120 

Art. XI. — Saccoglottis, Recent and Fossil; by E. W. Berry, 127 

Art. XII. — A Crossotheca from the Rhode Island Carbon- 
iferous; by E. M. Round, 131 

Art. XIII. — A Fossil Dogwood Flower; by F. H. Knowlton, 136 

Art. XIY. — Intrusive Rocks of the Portsmouth Basin, Maine 

and New Hampshire; by A. Wandke, 139 

Art. XV. — Babingtonite from the Contact Metamorphic 
Deposits of the Yakuki Mine, Province Iwaki, Japan; 
by M. Watanabe, .". 159 

Art. XVI. — A Tillite-like Conglomerate in the "Eocambrian" 

Sparagmite of Southern Norway; by O. Holtedahl, . . . 165 

Obituary.— A. Gr. Mayor, 173. 

Library, U. S. Nat. Museum. 


Established by BENJAMIN SILLIMAN in 1818. 




Editor: EDWARD S. DANA. 



op Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 

of Washington. 



No. 21 -SEPTEMBER, 1922. 

19 22. V^ 


Published monthly. Six dollars per year, in advance. $6.40 to countries in the 
Postal Union ; $6.25 to Canada. Single numbers 50 cents; No. 271, one dollar. 

Entered as second-class matter at the Post Office at New Haven, Conn., under the Act 
of March 3, 1879. 

These Are Wiley Books 

Elements of Optical Minerals 

An Introduction to Microscopic Petrography. 

By N. H. WlNCHELL and A. N. WlNCHELL. Second Edition. En- 
tirely Rewritten and Much Enlarged by ALEXANDER N. WlNCHELL, Doct. 
Univ. of Paris, Professor of Mineralogy and Petrology, University of Wisconsin. 
Part I. — Principles and Methods. 

The treatment given in this first Part of the second edition, taking up 
Principles and Methods, is adapted to the microscopic study of Minerals, both 
in thin sections and in powdered form immersed in liquids, 
the latter method permitting an accurate quantitative measurement of the index 
of refraction. 

(Now going through the press — ready October 1st.) 

The Famous 'Dana' Mineralogies 

Textbook of Mineralogy System of Mineralogy 

With an Extended Treatise on 
Crystallography and Physical Miner- 
alogy. Third Edition, Ee vised and 
Enlarged by WILLIAM E. FORD. 
720 pages, 6 by 9, 1050 figures, $5.00 

Manual of Mineralogy 

Thirteenth Edition, entirely Ee- 
vised and Eewritten by WILLIAM E. 

460 pages, 6 by 9, 357 figures, 10 plates. 
Cloth, $3.00. Flexible binding, $3.50. 

(Descriptive Mineralogy.) Sixth 
Edition, with Appendices I and II, 
completing the work to 1909. 
1323 pages, 6fby 10, over 1400 figures; 
half leather, $15.00. 

(Third Appendix to "System," 
completing the work to 1915, by WlL= 
LI AM E. FORD, issued in separateform. 
87 pages, 6f by 10, cloth, $2.00.) 

Minerals and How to 
Study Them 

A Book for Beginners in Miner- 
380 pages, 5 by 7, 319 figures, $2.00. 

Send for these books on our liberal Free Examination terms — TODAY. 


432 Fourth Avenue 
New York 


Montreal, Quebec 
Renouf Publishing Company 
AJS 9.22 




Aet. XVII. — The Determination of the Space Group of a 
Cubic Crystal; by Ralph W. G. Wyckoff. 


The theory of space groups presumably defines all of 
the ways in which elements of symmetry may be distri- 
buted in space so that their aggregates will possess 
crystallographic symmetry. 1 A knowledge of the space 
group to which a particular crystal should be assigned 
thus describes completely its characteristics of symme- 
try, and forms thereby one of the principal goals of 
descriptive crystallography. On the basis of direct 
experimental evidence it has hitherto been impossible to 
carry crystallographic description so far ; only in a few 
isolated cases could the appropriate space groups be 
inferred. 2 Use of the diffraction effects resulting from 
the action of X-rays upon crystals offers, however, the 
opportunity in many cases of determining experimentally 
the space group corresponding to a crystal. 

The crystal symmetry which is deduced by the use of 
X-ray methods of study is the symmetry of the arrange- 
ment of the atoms of which the crystal is composed. The 
identification of this internal symmetry with the external 
crystal symmetry, obtained from studies of face-develop- 
ment and the like, requires an assumption equivalent to 
one that states that the external symmetry of a crystal is 
consistent with the arrangement of its constituent atoms. 
Not only does it appear natural to relate them thus but 
the generally satisfactory agreement between the exter- 
nal symmetries and the symmetries of the crystal models 
of those crystals whose structures have thus far been 
studied with X-rays points to the correctness of this 

1 A. Schoenflies, Krystallsysteme und Krystallstructur, Leipzig, 189L 

2 For instance, C. Viola, Z. Kryst., 28, 225, 1897; L. Sohncke and E. Fed- 
erov have made similar assignments. 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 21. — September, 1922. 

176 R. W. G. Wyckoff — Determination of the 

The probable space groups corresponding to a few 
crystals have already been determined by showing that 
the structures assigned to them by the X-ray studies can 
be deduced from certain particular space groups. 3 It is 
the intention of the present paper, however, to show, by 
taking the cubic crystals as the simplest examples, how 
the space groups of many crystals can be uniquely deter- 
mined in advance of a complete elucidation of their struc- 
tures and to state criteria which serve to distinguish 
between the various cubic space groups where such a 
distinction is possible. Not only is such a knowledge of 
the space group of a crystal an ultimate aim of formal 
crystallography but it may be of great value in the 
problem of crystal structure study itself. 

In the crystals whose structures have been determined 
all, or nearly all, of the atoms of which they are composed 
have been found to occupy positions within the unit cells 
whose coordinate values are limited and denned by sym- 
metry considerations (the corners, center, centers of the 
edges and of the faces of a unit cube are such positions). 
Such very special structures can usually be deduced from 
more than one space group. Most cubic crystals, how- 
ever, have one or more of their constituent atoms in posi- 
tions so general that the symmetry requirements permit 
their x, y and z coordinates to have any values. Physical 
data concerning the mechanism of the scattering of 
X-rays by atoms are not yet sufficient for the complete 
determination of the structure of any crystal having 
one or more atoms in these general positions. A knowl- 
edge of the space group to which such a crystal should be 
assigned serves to determine the manner of arrangement 
of its atoms in many cases, even though the distances 
between some of these atoms cannot now be established 
with accuracy. 

X-ray criteria for distinguishing between the different 
space groups are most simply and satisfactorily deduced 
for those crystals having some atoms of appreciable 
scattering power in general positions. For this reason 
and also because a knowledge of the space group of such 
crystals is valuable to the crystal analyst, the discussion 
which follows will be limited to cubic crystals having one 
or more atoms in general positions. Similar criteria 

3 A. Johnsen, Physikal. Z., 16, 269, 1915. 

Space Group of a Cubic Crystal. 177 

have been established for other than cubic crystals, 
although their application to specific instances is not so 
straightforward. Some discrimination among the 
special cases where one or more of the coordinates of 
position are defined by considerations of symmetry can 
likewise be made. 

Means of distinguishing between cubic space groups as 
an aid to studies of crystal structure by reflection spec- 
trum observations of the relative spacings against the 
(100), (110) and (111) faces have already been given. 4 
For a variety of reasons, however, this method is of little 
•certain value in actual practice. 

A study of the Laue photographs taken in a single 
direction through a crystal distinguishes as far as possi- 
ble between the various space groups. Because of the 
much larger mass of data with which they deal, space 
group determinations based upon Laue photographic 
studies are not' open to the same measure of uncertainty 
as those derived from reflection spectrum measurements. 

It will be seen from the criteria to be discussed that 
many of the space groups give diffraction effects which 
are different from those given by any other space group 
and thus a method is established for deducing completely 
and uniquely the (internal) symmetry of a crystal with- 
out recourse to methods of studying external symmetry, 
such as those of face-development and etch-figure forma- 

These space group criteria have already been used 
upon a number of crystals of rather complicated chemical 
compositions. Of these, studies of nickel nitrate hexam- 
monate 5 and of sodium hydrogen acetate 6 have either 
been published or are in the course of publication; the 
determination of the symmetry and structure of zinc 
bromate hexahydrate, published elsewhere in this 
Journal, 7 has been written primarily to serve as an illus- 
tration of the application of these criteria. 

Methods of Distinguishing between the Cubic Space Groups. 
The general characteristics of the diffraction effects to 

4 P. Xiggli, Geometrische Kiystallographie des Discontirimims, p. 492, 
Leipzig, 1919. 

5 Ralph W. G. Wvekoff, Jour. Am. Chem. Soc, June, 1922. 

6 Ralph W. G. Wvekoff, see the third article in this number of this Journal. 

7 See the following article in this number. 

178 R. W. G. Wychoff — Determination of the 

be obtained from atoms arranged according to the 
general positions of any one of the space groups can be 
readily calculated with the aid of the customary intensity 
expression : 

I OC / (^\[A*+B*l where 

A = ^d [<r m co&%irn(hx m +ki/ m -\-lz m )], [1] 


and B is a similar sine term. In this expression 8 1 is the 
intensity and n is the order of reflection from a plane 
whose Miller indices are (hM) ; x m ,y m ,z m are the coordi- 
nate positions of each of the m atoms (within the unit) 
over which the summation is to be extended and o- m is the 
scattering power of the atom m. The value of 


where d hkl is the spacing of the plane (hhl), need not be 
evaluated for the present purposes. If these diffraction 
effects are calculated for each of the space groups it 
is found that for some of them reflections from certain 
classes of planes in some orders will be entirely absent. 
Such a complete absence of definite classes of pLanes, 
different for different space groups, makes it possible to 
distinguish between these groups. 

As an example of the mode of procedure a common 
space group, T h 6 , the sixth group having paramorphic 
hemihedral (pyritohedral) symmetry, will be considered 
in detail. The coordinates of the most generally placed 
equivalent points within a unit cube for this space group 

xyz; x + %, \—y , z; x,y+i, \—z\ i—x, y, z+\ 
zxy; z, a?+£, |— y; %—z, x, y+$; z + i, \—x, y 
yzx; %—y, z, x + %; y + i, £— «, as; y, £+^, \—x 
xyz; \—x, y+i, z\ x, $ — y, z+i; x+%, y, \—z 
zxy; z, i—x, y+i; z+i, x, \ — y; \—z, x+\, y 
yzx; y+i, z, \—x\ f— y, z + $, x) y, \—z, x+\. 

Taking for the present purposes the scattering power 
(o-) of atoms in these general positions as unity and divid- 

8 Kalph W. G. Wyckoff, this Journal, 50, 317, 1920. 

Space Group of a Cubic Crystal. 179 

ing all reflecting planes into three classes having indices 
that are (1) two even and one odd, (2) two odd and one 
even and (3) all odd, the A and B terms of expression [1] 
are found to be as follows for the first order region of 
the spectrum (which can be distinguished in Laue photo- 
graphic data from cubic crystals without any uncer- 
tainty) : 

(1) "When the indices are two even and one odd, and 
p, c[, and r are any integers : B = 0, and 

A=2cos 27r[2px+2qy+(2r+\)z] + 2cos 27r[-2px+2qy-(2r+l)z + $] 

+ 2 " I2pz+2qx+(2r+l)y] +2 

+ 2 " [2py+2qz+(2r+l)x] +2 

+ 2 " [2px-2qy-(2r+\)z\ +2 

+ 2. " [2pz-2qx-(2r+l)y] +2 

+ 2 " [2py-2qz-(2r+X)x : ] +2 

[-2pz + 2qx-(2r+l)y+1 i ] 

(2) When the indices are two odd and one even: B = 

A=2cos 27r[223x+(2q+l)y+(2r+l)z] +2cos 27r\2px-(2q+l)y-(2r+l)z + l] 
+ 2 " I2pz+(2q+'i)x + (2r + \)ii] +2 " \2pz-{2q+\)x-{2r+\)y+\] 
+ 2 " [2py+(2q+l)z+(2r + l)x] +2 " [2py-{2q+l)z-(2r+l)x + \] 
-1-2 " [-2px+(2q+l)y-(2r+l)z] +2 " [-2px-(2q+l)y+(2r+i)z + $] 
[-2pz+(2q+l)x-(2r+l)y] +2" [-2pz-(2q + \)x+(2r+l)y-±] 
[-2py + (2q+l)z-(2r+l)x] +2 " [-2py-(2q+l)z+(2r+l)x + i] 


(3) When the indices "are all odd : B = 0, and 

A = 2cos 2-[(2p + l)x+(2q + l)y+(2r+l)z) +2cos 2n[-(2p+l)x+(2q+l)y-(2r+l)z] 

+ 2 " $(2p+l)z+(2q+l)x+(2r+l)y] +2 
\(2p + l)y+(2q+l)z+(2r+\)x] +2 

2 " f(9,-n-L--\Yy>—<0 r ,i-\\,.(O- i ,_ L .-\\~-\ iO 


l(2p + l)x-(2q+l)y-(2r+l)z-] 


[(2p+l)z-(2q+l)x-(2r+l)y] +2 
+ 2 " [(2p+l)y-{2q+l)z-(2r+l)x] +2 

[-(2p + l)x-(2q+l)y+@r+l)z] 


It is thus seen that in general all three groups will 
appear in the first order region of the spectrum. The 
following procedure will, however, serve to determine 
whether there may not be classes of planes within these 
groups which will show a different behavior. 
cos27r(a)= — cos2tt( / 8), when a=(±/3±^) 

Consequently any and all values of p, q, and r which 
will make A = for the region n = 1 can be found by 
equating the revolutions of the first term of A, [Zpx+Zqy 
+ (2r+l)s], to the revolutions of each of the other terms 
of A plus i (and any integer s) and solving the resultant 

180 R. W. G. Wyckoff — Determination of the 

expressions for any possible integral valnes of p, q, and r. 

2px + 2qy + (2r+l)z= + [2pz + 2qx+(2r+\)y + i+s~], 
2px+2qy + (2r+l)z = -^[2py+2qz+{2r + l)x+^ + s], 

2px + 2qy + (2r+l)z = + [2px—2qy-(2r+l)z + $+s], 


It is readily shown that all of the solutions to any of 
these equations are comprehended by making p = and 
letting q and r have any values. A similar set of equa- 
tions can be set up for the second and for each succeeding 
term of A and values of p, q and r which will make A = 
can be selected from those solutions which are common to 
all of these sets of equations. Because of the simplicity 
of these expressions this detailed procedure can be mater- 
ially shortened in actual practice. 

If the B term were not invariably equal to zero, a simi- 
lar procedure would have to be followed and solutions 
common to it and to the A term chosen. Since 

sin 27r(a) = — sin 2tt(/?) 

both when a = — (3 and when a— (/?±-|), two sets of 
expressions somewhat different from those of the A 
terms must be established. 

The carrying out of this procedure for each of the three 
groups of planes shows that 

(1) When the indices are two even and one odd, A = 
0if2p = h,2q = k = 0, (2r + 1) = I; 

(2) "When the indices are two odd and one even, A = 
if 2p = h = 0, (2q + 1) = k, (2r + 1) = I; ' 

(3) When the indices are all odd, A is never equal to 
zero. This absence in the first order of planes of the 
class hol, where h is even, and of the class Okl, where both 
k and I are odd, will then be a universal characteristic of 
the def raction effects from all crystals having the symme- 
try of the space group T h 6 . Since a further study of all 
of the space groups shows that there is no other one for 
which these classes of planes and no others are absent, a 
unique method is thus provided for determining from a 
study of its Laue photographs whether or not a crystal 
has the symmetry of the space group T h 6 . 

By extending the typical treatment applied here to 
each of the cubic space groups, a series of criteria can be 

Space Group of a Cubic Crystal. 181 

established for distinguishing in so far as is possible 
between them. Since in tetartohedral and paramorphic 
hemihedral (pyritohedral) crystals the plane likl belongs 
to a different form from kill (for instance 041 is distinct 
from 401) the Laue photographs to which these crystals 
give rise will possess less symmetry than photographs of 
those belonging to one of the other classes of cubic sym- 
metry. On the basis of an observed hemihedry or holo- 
hedry in the symmetry of the Laue photographs, crystals 
of the classes T or T h can always be distinguished from 
crystals of the classes T d ,0 or O h . 

Crystals corresponding to space groups based upon r c> 
the simple cubic lattice, will in general like T h 6 give 
reflections in the first order from planes of all three 
groups ; those based upon the face-centered lattice, r c ', 
will give first (or any odd numbered) order reflections 
only from planes all of whose indices are odd ; and those 
developed from the body-centered lattice, r c " ? will reflect 
in the first (or any odd numbered) order only planes 
having two indices that are odd and one even. Upon the 
observed symmetry of the corresponding Laue photo- 
graphs and the nature of the underlying lattice, the cubic 
space groups can be given the following preliminary 
arrangement : 

r c : all kinds of planes in all orders, 
Hemihedral Laue Photographs : 

Tl T4 m i m 2 m « , 
J - 1 - ? -■- h 7 J- h J - 1 - li f 

Holoheclral Laue Photographs : 

iy : only all odd planes in odd orders, 
Hemihedral Laue Photographs : 

T2 rp 3 m 4 . 

Holohedral Laue Photographs : 
T d 2 ,T/,0 3 ,OSO h 5 ,O h 6 ,Oj,O h 8 . 
r c " : only two odd and one even planes in odd orders, 
Hemihedral Laue Photographs : 

Holohedral Laue Photographs : 

T^T d 6 ,0 5 ,0 8 ,O h 9 ,O h 10 . 

182 R. W. G. Wychoff — Determination of the 

Distinctions between most of the space groups of any- 
one of these divisions are possible in the first order region 
of the spectrum (see page 183) . By calculating the A and 
B terms of the intensity expression [1] for the second- 
order region for various space groups, using the same 
procedure previously employed, a few more distinctions 
between space groups can also be made*. A final classi- 
fication of all of the cubic space groups on the basis of the 
difTr action effects produced by corresponding crystals 
can be written as follows. In this table indistinguishable 
space groups are placed together on one line. 

r c : 



T h 2 

T d 4 ,O h ! 

O h 2 ; 


r '• 

T 4 ; \ 




rp2 rr\ s . \ 

' rr/V r second order region 

ih ; ) 

T d 2 ,0 3 , ) second order region, O h 6 ; \ second or 
O*, [ uncertain, O h 7 ; [ region 

T 6 O 6 • ) 

'/-VV r second order region 


T 3 ,T 5 ,T h 5 ; 

T h 7 ; . 

T d 3 ,0 5 ; ) second order region, O h 9 ; 
O 8 ; f uncertain, 

T d 6 ; 
O h 10 . 

Except where definitely stated as lying in the second- 
order region, the distinguishing characteristics are to be 
understood as being first-order effects. If it is assumed, 
as may or may not be the case, that studies of face 
development as commonly carried out upon crystals 
invariably indicate the symmetry of the arrangement of 
their atoms, then it will be seen that means are at hand 
for distinguishing between all of the various space 
groups except the two pairs T 3 and T 5 and O 6 and O 7 . 
Those distinctions which involve the absence of planes 

Space Group of a Cubic Crystal. 183 

of a single form (the {100} planes) can, however, be used 
if at all with only the greatest caution because it may 
readily happen that the scattering powers and relative 
positions of different atoms in a crystal will be such as to 
make the reflections from planes of this form so weak as 
not to be observed under the ordinary conditions of 
experimentation. Distinctions of this sort have conse- 
quently been designated as uncertain in the preceding 

The characteristics serving to distinguish between 
each of the space groups can be stated as follows : 

Space Groups based upon a Simple Cubic Lattice, in 
general all three kinds of planes appearing in all orders : 

Hemihedral Photographs : 

T 1 and TV : Xo classes of planes absent ; 

T 4 : Planes of the form {100} absent in odd orders ; 

T h 2 : Planes of the form {Old), where k and I are 
one even and the other odd, are absent in odd orders ; 

T h 6 : Planes of the form {liOl), where k is even and 
I is odd, and of the form {Okl}, where h and I are both 
odd, are absent in the odd orders. 

Holohedral Photographs : 

Td 1 , O 1 , and (V : No classes of planes absent. 

T d 4 and O h 3 : Planes of the form {hM}, where h = 
± ~k and either li is even and I is odd or both h and I are 
odd, are absent in odd orders ; 

O 2 : Planes of the form {100} are absent in odd 
orders ; 

O 6 and O 7 : Planes of the form {100} are absent in 
all but the fourth, eighth, etc., orders ; 

O h 2 : Planes of the form {old), where Jc is even and 
I is odd, and of the form {7*7?/}, where either h is even and 
I odd or both h and I are odd, are absent in odd orders. 

O h 4 : Planes of the form {Old}, where h is even and 
I is odd, are absent in the odd orders. 

Space Groups based upon a Face Centered Cubic 
Lattice, in general planes having all odd indices appear- 
ing in odd orders : 

184 R. W. G. Wyckoff — Determination of the 

Hemihedral Photographs : 

T 2 , T h 3 , and T h 4 : No classes of all odd planes are 
absent in the first order. Second order reflections are 
absent from T h 4 for planes of the forms {Ohl} and {hOl}, 
where h is even and I is odd. 

Holohedral Photographs : 

T d 2 , O 3 , 0% O h 5 , and O h 7 : No classes of all odd 
planes are absent in odd orders. From O 4 only fourth, 
eighth, etc., orders from planes of the form {100} are 
present. Second order reflections from O h 7 are absent 
for planes of the form {Old} where k is even and I is odd; 

T d 5 , O h 6 , and O h 8 : Planes with all odd indices are 
present in odd orders except those of the form {KM}, 
where h = ± k. Second order reflections from O h 8 of 
planes of the form {Ohl}, where k is even and I odd, are 

Space Groups based npon a Body Centered Cubic 
Lattice, in general planes having two odd and one even 
indices appearing in odd orders : 

Hemihedral Photographs: 

T 3 , T 5 , and T h 5 : No classes of two odd and one 
even planes are absent ; 

T h 7 : Planes of the form {Okl}, where k and I are 
both odd, are absent from odd orders. 

Holohedal Photographs : 

T d 3 , O 5 , O 8 and O h ° : No classes of two odd and one even 
planes are absent in the first order. Reflections in the 
second order from planes of the form {100} are absent 
for O 8 . 

T d G : Planes of the forms {Ml} where \h is even and I 
odd are absent in odd orders. In the second order, planes 
of the forms {hhl}, where either h is even and I odd, or 
both h and I are odd, are absent. 

O h 10 : Planes of the form {Okl}, where k and I are odd, 
and of the form {Ml}, where \h is even and I is odd, are 
absent in the first order. Reflections in the second order 
are absent for planes of the forms {hhl} where either 
h is even and I odd or both h and I are odd. 

Space Group of a Cubic Crystal. 185 

It will be observed that a number of space groups give 
rise to diffraction effects which are different from those 
resulting from anv other space group. These unique 
space groups are * T h 2 ,T h 6 ,O h 2 ,O h 4 ,T h 4 ,(V ,O h s ,T h 7 ,T d 6 ,O h 10 ; 
and, less certainly because they depend upon the presence 
or absence of planes of the single form 100; T 4 ,0 2 ,0 6 and 
7 ,O,0 s . The symmetry of a crystal corresponding to 
any one of these space groups can consequently be deter- 
mined with complete certainty without any reference to 
face development, etch-figure symmetry or any other of 
the customary methods of crystallography. A possible 
experimental method is thus furnished for finding out 
what relations exist between the symmetry assigned to a 
crystal by studies of its external appearance and the sym- 
metry of the arrangement of its atoms. 

The experimental establishment of the space group of 
a particular crystal is simple and can be carried out by 
the procedure common in crystal structure determination 
of taking one or more Laue photographs about some con- 
venient orientation, determining the indices of the various 
diffraction spots by the usual methods of projection and 
finding the wave-lengths of the X-rays producing these 
spots with the aid of a measurement of the dimensions of 
the unit cell through a reflection spectrum measurement 
from some convenient crystal face. 9 Then if the voltage 
applied to the X-ray tube in producing the photographs 
is known, the range of the spectrum in which there will be 
only first-order reflections can immediately be told. It 
happens that in all cases where it is necessary to go to the 
second-order region to distinguish between space groups, 
first-order reflections from the planes involved are also 
missing. No ambiguity is therefore introduced concern- 
ing the order of reflection of diffraction spots lying in the 
region of strong second-order effects. Reflections from 
faces of the form {100} are as a rule more readily 
obtained by reflection spectrum measurements than from 
Laue photographs. It is desirable to reemphasize that, 
for the reasons already given, only the appearance and 
not the absence of {100} reflections can necessarily 
furnish conclusive evidence upon which to base assign- 
ments to particular space groups. 

In certain of the cases where the diffraction data are 

9 Ralph W. G. Wyckoff, this Journal, 50, 317, 1920. 

186 R. W. G. Wyckoff — Determination of the 

insufficient, a knowledge of the numbers of chemical 
molecules to be associated with the unit cube, such as 
arises immediately from the density and the dimension of 
the unit, can be of service. For instance suppose that 
the diffraction data from a certain crystal assigned it to 
the two indistinguishable space groups T 3 and T 5 , and 
suppose that the determination of the number of chemical 
molecules within the unit cube requires but two chemi- 
cally like atoms within the unit cell, then since only T 3 
contains as a special case two equivalent positions, the 
crystal may be assigned to it, rather than to T 5 . In view 
of the present lack of definite knowledge as to what it is 
that conditions chemical equivalence in the crystalline 
state, such information must obviously be used with great 

There naturally arises a question of whether even with 
atoms in the most general equivalent positions coordinate 
values may not exist such that the diffraction results may 
simulate those corresponding to some space group other 
than the one to which it really belongs. Any such coordi- 
nates for a space group can readily be found by practi- 
cally the same procedure which has already been 
employed in determining the reflection characteristics of 
planes in different orders. In this process, however, the 
sets of equations are to be solved for x, y and z rather 
than for h, k and I. 

The space group T h 6 will again serve as an illustration. 
The previously established set of equations, [2], must 
now be solved for x, y and z which can have any values 
between zero and unity, including the former, instead of 
for integral values of p, q and r. For the present purpose 
care must of course be taken to avoid such values of x, y 
and z as yield special cases with fewer than the maximum 
number of equivalent positions within the unit cell. By 
solving these sets of expressions in a manner analogous 
to that previously used it can be shown for instance that 
when x = u, y = and z = 0, or when x = -J, y = \ and 
£ — 0, only all odd planes are to be found in the first 
order region. "When attempting to ascertain the space 
group to which a crystal should be assigned, it is impor- 
tant to take into consideration the possibility of atoms 
occupying exactly or nearly such positions as these. It 
must likewise be borne in mind that atoms in special posi- 

Space Group of a Cubic Crystal. 187 

tions may not give diffraction effects in certain orders ; 
so that in cases where most of the heavy atoms are in such 
positions, the characteristic effects upon which a choice of 
space groups is made may some of them be relatively 
weak. Especially in view of this possibility of atoms 
occupying in some instances coordinate positions which 
may alter the qualitative character of the resulting dif- 
fraction effects, it is necessary to emphasize the fact that 
though these criteria are not ambiguous when used 
properly they cannot be applied blindly. 


Criteria, which are valid for crystals which have any 
atoms of appreciable scattering power in general posi- 
tions, are established for determining from studies of 
Laue photographs the space group to which a cubic 
crystal should be assigned. This knowledge is of value 
to the crystal analyst because it is thus possible to tell 
how the atoms in many chemically complicated crystals 
are arranged, even though existing methods are not suffi- 
cient to locate these atoms with accuracy, and because an 
assignment of a crystal to a particular space group 
defines completely the positions of all of its elements of 
symmetry. Many of the space groups give diffraction 
effects which are different from those given by any other 
groups and hence a method is provided, in the cases of 
crystals assignable to any of these unique space groups, 
of denning completely crystal symmetry without making 
use of the older methods such as face development and 
the like. 

Geophysical Laboratory, 
April, 1922. 

188 R. W. G. Wyckoff — Symmetry and Crystal 

Aet. XVIII. — The Symmetry and Crystal Structure of 
Zinc Bromate Hexahydrate, Zn(Br0 3 ) 2 .6H 2 0; by 
Ralph W. Gr. Wyckoff. 1 

[Contribution from the Gates Chemical Laboratory of the California Insti- 
tute of Technology, No. 16.] 


This paper has the two-fold purpose of adding confir- 
mation to the previously assigned structure of nickel 
nitrate hexammonate 2 by the study of an analogous 
compound and, more especially, it is intended to be an 
illustration of the application of those criteria for distin- 
guishing between the cubic spa'ce groups which are 
described in the preceding article^ 

Excellent crystals of Zn(Br0 3 j 2 .6H 2 0, mostly octahe- 
dral in habit, are formed from solutions both by slow 
cooling and by slow evaporation. The crystals that grow 
from a cooling solution usually exhibit a slight anoma- 
lous double refraction with sectoring. 4 The Laue photo- 
graphs to which these crystals give rise do not, however, 
show any anomalous effects. Completely isotropic speci- 
mens are obtained by gradual evaporation. 

Zn(Br0 3 ) 2 .6H 2 is one of a group of isomorphous 
crystals to which belong the chlorates of nickel, cobalt 
and probably copper, and the bromates of nickel, cobalt 
and magnesium. 5 

The Structure of Zinc Bromate Hexahydrate. 

A reflection photograph from the octahedral face 
combined with an estimation of the density of the salt 
indicates that four chemical molecules are to be asso- 
ciated with the unit cube. 6 The length of the side of this 
unit was found to be 10.31 A.U. (10.31 X 10- 8 cm.). 

Laue photographs were prepared through both octahe- 

1 Member of the Staff of the Geophysical Laboratory of the Carnegie 
Institution of Washington. 

2 Ealph W. G. Wyckoff, Jour. Am. Chem. Soc, June, 1922. 

3 See page 175 of this Journal. 

4 Marbach, Poggendorffs Ann. d. Phys., 99, 465, 1856. 
5 P. Groth, Chemische Krystallographie, II, p. 112, Leipzig, 1908. . 
fl Ealph W. G. Wyckoff, Jour. Am. Chem. Soc, 42, 1100, 1920 ; Ealph W. G. 
Wyckoff and Eugen Posnjak, ibid. 43, 2292, 1921. 

Structure of Zinc Br ornate Hexahydrate. 189 

dral and cube faces. These showed clearly an absence 
of planes of symmetry; hence it is evident that the 
symmetry of the arrangement of the atoms of this crystal 
is either tetartohedral or paramorphic hemihedral (pyri- 
tohedral). Interpretation of these photographs in the 
usual manner 7 showed that in general planes of all sorts 
appear in the first order region. The fundamental 
lattice must consequently be the simple cubic lattice. 
There are four zinc atoms within the unit, and it is both 
natural and in accord with previous experience to con- 
sider them equivalent. If, merely to serve as a starting 
point for considering the various possible space groups, 
this assumption of the equivalence of the zinc atoms is 
made, we find that there are four tetartohedral and para- 
morphic space groups built upon a simple cubic lattice 
which have as special cases four equivalent positions 
within the unit, namely the groups T^T^TVSTV 3 . 

An inspection of the criteria for distinguishing 
between these space groups (see the preceding article) 
suggests the investigation of those planes having one of 
the indices zero. Some data for first order reflections of 
such planes from a Laue photograph with the X-rays 
roughlv normal to an octahedral face are given in 
Table I. 


I. Laue Photographic Data. 

From a Plate taken 

with the X-rays roughly 

normal to (111) 

Appearing Planes. 

Indices of plane 

Wave Length 

Form of plane 


0.480 A. U. 
































7 Kalph W. G. Wyckoff, this Journal 50, 317, 1920. 

190 R. W. G. Wyckoff — Symmetry and Crystal 

Absent Planes. 














































From this table it will be seen that the only planes of 
this type which appear in the first order region are of the 
forms {JtQl}, where h is even and I is odd ; it is also appar- 
ent that many planes of the forms {Ohl} and of the forms 
{Okl}, where both k and I are odd, were in suitable posi- 
tions to reflect but did not do so. Results in complete 
agreement with these data and from planes of still differ- 
ent forms are obtained from the interpretation of a 
photograph taken with the X-rays approximately normal 
to a cube face. In comparing the data obtained from two 
different Laue photographs of either a tetartohedral or 
paramorphic crystal, it must of course be remembered 
to choose the H and K axes in the same way in both cases ; 
this is readily accomplished by observing planes of two 
forms {hkl} and {kill} which show marked differences in 
reflecting power and are common to the-two photographs 
to be compared. 

The data recorded in Table I are seen to be in entire 
accord with the criteria which determine the space group 
T h 6 . Since these criteria uniquely distinguish this space 
group from every other group, it is evident that the 
symmetry of crystals of zinc bromate hexahydrate is that 
of T h 6 . From this knowledge of the corresponding space 
group and the fact that four chemical molecules are to be 

Structure of Zinc Bromate Hexahydrate. 191 

associated with the unit cube, the manner of arrange- 
ment of the atoms of zinc bromate hexahydrate is defi- 
nitely determined to be as follows : 

Zinc atoms : Arrangement 4b, 

000: U0; OH; £0£. 

Bromine atoms : Arrangement 8h, 

uuu; u+i, | — u, it: u. it4-£, i—u; |— v, u, U + -J-; 
uiiu: i — u y w+ij u; u, $ — v.. u + i; u+i, u, 4 — u. 

Bromate oxygen atoms : General positions, 

xyz; x+±, l—y, z: x, y+i, \—z; \—x, y, z+\; 
zxy; z, x+±, $—y; \—z, x, y+i; z-\-\ % \— as, y; 
yzx; % — y, z, « + £; y+k, \—z, x; y, z+$, i—x; 
xyz; i—x, y+$, z; x, -k—y, z+i; x+l, y, \—z; 
zxy; z y \—x, y+i: z+i, k, i— y; %—z, •x'+i, V: 
yzx; y + %, z, ±—x; \—y, z + \, x; y, \—z, x+\. 

Water oxygen atoms : General positions with different 
values of x, y and z. 

Hydrogen atoms : Two sets of general positions. The 
coordinates and terminology are taken from the writer's 
book entitled "An Analytical Representation of the 
Theory of Space Groups" which is shortly to be pub- 
lished by the Carnegie Institution of Washington. 

An inspection of the special cases of the space group 
T h 6 shows that the arrangement outlined above is the 
only reasonable one for the atoms of zinc bromate hexa- 
hydrate, since any other would string out sets of eight 
equivalent atoms along the body diagonals of the unit 
cube. Though it is impossible at the present time to 
obtain the positions of other than the zinc atoms, it is 
probable that the value of u, the parameter defining the 
bromine atoms, is in the neighborhood of -} and that the 
values of x, y and z for the bromate oxygen atoms are 
such as to cluster these atoms more or less closely about 
the bromine atoms. It will also be observed that the 
groups of atoms constituting the water molecules cannot 
be unequally distributed between the metal atoms and the 
bromate groups, but must all be related in exactly the 
same manner to the zinc atoms (or to the bromate 
groups). This distribution is in accord with that found 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 21. — September, 1fl22. 

192 R. W. G. Wyckoff — Zinc Br ornate Hexahydrate. 

for the ammonia groups in the hexammonate nickel 
halides s and in nickel nitrate hexammonate. 9 


From a study of the Laue photographs to which 
crystals of zinc bromate hexahydrate give rise, it is 
shown that they must have the symmetry of the space 
group T h 6 . Though it is impossible to determine the 
positions of the atoms in this crystal, such knowledge of 
the underlying space group defines uniquely the manner 
of arrangement of its atoms. The length of the side of 
the unit cube which contains four chemical molecules is 
found to be 10.31 A.U. 

Pasadena, California, 
May, 1922. 

8 Kalph W. G. Wyckoff, Jour. Am. Chem. Soc, June, 1922. 

9 Idem. ibid. 

R. W. G. Wyckoff — Sodium Hydrogen Acetate. 193 

Art. XIX. — On the Symmetry and Crystal Structure of 
Sodium Hydrogen Acetate, NaH(C 2 H 3 2 )2; by Ealph 
W. G. Wyckoff. 1 

[Contribution from the Gates Chemical Laboratory of the California Insti- 
tute of Technology, No. 17.] 


This study of the structure of sodium hydrogen acetate 
was undertaken in the attempt to find the arrangement 
of the atoms in some organic compound. 

Sodium acid acetate was prepared by the long con- 
tinued digestion of fused sodium acetate, glacial acetic 
acid and acetic anhydride in sufficient quantity to remove 
the small amounts of water which may be present. 2 It 
was thus obtained in cubes which under the polarizing 
microscope prove to be isotropic. Assignment to a par- 
ticular class of symmetry on the basis of the ordinary 
crystallographic evidence has never been made. 

The Study of the Structure of Sodium Hydrogen Acetate. 

Comparison reflection .spectra from the (100) face of 
calcite and the (100) face of sodium acid acetate showed 
three orders of reflection for the latter which stood in the 
ratio of 2:3:4 (experimental conditions did not permit 
of the first order registering itself). Measurements 
upon these photographs combined with the density, p = 
1.402, as determined by a flotation Westphal balance 
method, gives the mean value 3.07 for the ratio m/n z ,m 
being the number of chemical molecules within the unit 
cube and n the order of the reflection. There are thus 
either three or 24 chemical molecules of the composition 
NaH(C 2 H 3 02)2 within the unit. The length of the side 
of the cube having three molecules within it, as deter- 
mined by these same measurements, is 7.98 9 A.U. (7.98 9 X 
10- 8 cm.)! 

Several Laue photographs taken with the X-rays 
nearly normal to the cube face of crystals from two dis- 

1 Member of the Staff of the Geophysical Laboratory of the Carnegie Insti- 
tution of Washington. 

2 The writer is under obligation to Prof. H. J. Lucas and to L. M. Kirk- 
patrick for some of these preparations. 

194: R. W. G. Wyckoff — Symmetry and Crystal 

tinct preparations were studied. Some pertinent data 
from the interpretation of one of these photographs are 
given in Table I. In this table the wave-lengths of the 

Table I. 

Laue Photographic 


Indices of plane 


Wave Length 



0.232 A. U 



















1 10 1 




f + 


2 13 3 



6 13 3 



7 10 1 




m — 






f . 


11 5 4 






3 14 3 






Note: — In this table spots having an intensity / are faint, those marked 
m are of medium intensity and those designated by s are amongst the 
strongest appearing upon the photograph. 

reflected X-rays have been calculated on the basis of a 
small unit containing three chemical molecules. The 
voltage applied to the X-ray tube during the making of 
these photographs was such that no reflections are ordi- 
narily present in wave lengths shorter than A = 0.24 A.U. 
Since, as reference to Table I shows, appreciable effects 
with values of nX as low as 0.180 are to be found upon 
the photographs, the correct unit must contain 24 and not 
three molecules. 

This result was sufficiently unexpected, especially in 
view of the fact that a simple structure containing three 

Structure of Sodium Hydrogen Acetate. 195 

molecules was not only possible on the grounds of symme- 
try but was chemically plausible, that it seemed worth 
while to obtain a direct spectrum observation of a reflec- 
tion in the first order from this larger unit. This could 
be done by studying the secondary spectra from a cube 
face reflection. If a spectrum is taken from a crystal 
face in the usual manner, not only is the reflection from 

Figure 1. 

this face observed upon the photographic plate, but reflec- 
tions from various other faces which are brought into 
position by the continuous rotation of the crystal during 
the course of the experiment will be registered at the 
same time at various angles to the principal spectrum. 
Such a composite spectrum was prepared by passing the 
X-rays through a section of a crystal of NaH(C 2 H 3 2 ) 2 
mounted so that one of the cubic axes was coincident 
with the axis of its rotation and hence lay in the plane of 
the slit of the spectrograph. A reproduction of this 
spectrum is shown in fig. 1. 

196 R. W. G. Wyckoff — Symmetry and Crystal 

The identification of the secondary spectra on this 
photograph can be carried ont with the aid of the gno- 
monic net which has previously been described. 4 The 
distance from the crystal to the photographic plate can 
if necessary be accurately calculated from the measure- 
ment of the principal spectrum (in this case the (100) 
reflection). It is then a simple matter to prepare a 
gnomonic ruler 4 for plotting on a radius of five centi- 
meters the projections of the various secondary spectra. 
By making the distance from the crystal to the plate 
exactly five centimeters, it would be possible to use 
directly the same gnomonic ruler which serves for Laue 
photographs; by making this distance 10 cm., as is more 
satisfactory, the same ruler can of course be used by 
dividing by two the readings of the scale giving distances 
from the central spot. While, during the course of the 
experiment, the crystal is rotated back and forth, the 
gnomonic projections of the various reflecting planes 
will travel along the hyperbolas of the gnomonic net if 
it is so placed that the zero degree hyperbola (a straight 
line) coincides in position with the principal spectrum. 
By rotating the mean projected positions of these- reflect- 
ing planes (which in the present instance form a simple 
square network of points) through the angle of rotation 
suffered by the crystal, portions of hyperbolas will be 
decribed which will pass through the gnomonic projec- 
tions of the observed reflections for those planes that can 
reflect for this particular setting of the crystal. By 
superposing, then, the gnomonic projection of the photo- 
graph over the series of hyperbolic paths obtained in this 
manner, it is possible to identify the various reflections 
upon the photograph. Considerable care must be taken 
in such a determination and some ambiguity may of 
necessity arise because it frequently happens that reflec- 
tions from different planes, especially planes in different 
orders of reflection, give effects at about the same posi- 
tions upon the plate. 

By making this sort of an interpretation of the trans- 
mission spectrum photograph from sodium hydrogen 
acetate, reflections from planes belonging to a number of 
forms, such as {161}, {721}, and {521}, were observed in 

4 H. Hilton, Min. Mag., 14, 18, 1904; Ralph W. G. Wyckoff, this Journal, 
50, 317, 1920. 

Structure of Sodium Hydrogen Acetate. 197 

the first order from the large unit containing 24 chemical 

The Laue photographs show clearly the absence of 
planes of symmetry. The crystals of this compound 
must then have either tetartohedral or paramorphic 
hemihedral (pyritohedral) symmetry. It was further 
observed that all of the planes giving reflections in the 
first order region on the basis of the large unit have 
two odd and one even indices. This points to an underly- 
ing body-centered lattice. 5 The four space groups 
T 3 , T 5 , T h 5 and T h 7 have the appropriate symmetry and 
are built upon r c ". Distinction between the first three 
of these is impossible upon the basis of the diffraction 
effects to which they give rise. Crystals corresponding 
T h 7 would give no reflections 5 in odd orders from planes 
of the forms {0M} y where both h and I are odd. Neither 
upon the Laue photographs nor upon the transmission 
spectrum photograph were any such planes found in odd 
orders, even though some were in suitable positions for 
reflection. This would make it necessary to assign to 
crystals of sodium acid acetate the symmetry of T h 7 . The 
unit is so large, however, that with moderate degrees of 
tilt from symmetrical Laue photographs the few planes 
which reflect in the first order region have complicated 
indices and are of weak intensity. Consequently in order 
to place this assignment of symmetry beyond any legiti- 
mate questioning, it would be desirable to study Laue 
photographs from crystals inclined farther from the 
symmetrical position than those here investigated. It 
did not, however, seem worth while to make these addi- 
tional experiments at the present time. 

Accepting this assignment to the space group T h 7 as 
correct, the general coordinate positions of the atoms in 
sodium acid acetate are defined. Depending upon what 
equivalence is assumed for the two acetate groups and 
for the atoms within these groups, all of the atoms of the 
crystal will be arranged according to either the 48 
generally equivalent positions 6 of T h 7 or the 24 equivalent 

5 Ralph W. G. Wyckoff , see the first article by the writer in this number 
of this Journal. 

These general positions are stated in abbreviated form by A. Schoen- 
flies, Krystallsysteme und Krystallstruktur, p. 551, Leipzig, 1891. Also in 
detail by P. Niggli, Geometrisehe Krystallographie des Discontinuums, p. 
368, Leipzig, 1919. 

198 R. W. G. Wyckoff — Sodium Hydrogen Acetate. 

positions of the special case 24e. It is of course impos- 
sible now to determine the positions of any of these 
atoms. The coordinates of both of these arrangements 
are given in the book by the writer entitled ' ' The Analy- 
tical Representation of the Theory of Space Groups'' 
which is shortly to be published by the Carnegie Institu- 
tion of Washington. 

When it is considered that such a relatively simple sub- 
stance as this sodium acid acetate has such a very compli- 
cated structure as it has here been shown to possess, it 
scarcely needs to be emphasized that any studies of the 
structures of organic compounds should be made with 
extreme care and in order to be of any value should make 
use of the most powerful diffraction methods now at our 


It is shown that the unit cell of sodium hydrogen 
acetate must contain 24 chemical molecules. The length 
of the side of this unit cube is found to be 15.98 A.U. 
The determination of the underlying space group as 
probably T h 7 defines the general manner of the arrange- 
ment of the atoms of this crystal, though it is impossible 
to obtain the positions of these atoms. A graphical 
method is outlined for identifying the planes causing the 
secondary spectra upon a reflection spectrum photo- 

Pasadena, California. 
May, 1922. 

W. A. Tarr—Cone-in-Cone. 199 

Akt. XX. — Cone-in-Cone; by W. A. Tabr. 1 

Cone-in-cone is a structural feature found in shales and 
rarely in coal. It is usually associated with concretions 
but not necessarily so. An occurrence of cone-in-cone in 
coal and its development in bands of calcite are such 
exceptions. The cone-in-cone structure consists of a 
series of cones within cones, adjacent cones uniting to 
form lenses or layers. When associated with concre- 
tions the cone-in-cone may occur on the upper or lower 
surface, or, more rarely, within the concretion. 

The structure was recognized and called cone-in-cone 
in the early part of the last century. It was not so called 
by all, however. Hildreth, in 1836 (see literature at end 
for all references), described it as a "fossil columnar 
Madrepore." It has also been called "cone-in-cone 
coral." The "German name for cone-in-cone is "tuten- 
mergel" and was given to it as early as 1823. 

Cone-in-cone has been described by Marsh, Sorby, New- 
berry, Jukes, Dawson, Daintree, Young, Sach, Garwood, 
Gresley, Broadhead, and many others. Their descrip- 
tions are all very similar, although their conclusions 
regarding the origin of cone-in-cone differ considerably. 

Probably the majority of investigators have regarded 
cone-in-cone as having been caused by pressure which 
was due (since the structure is so frequently seen in asso- 
ciation with concretions) to the expansion of concretions 
through growth. Thus, most cone-in-cone structures are 
regarded as of secondary origin. Some of those holding 
this view are Marsh, Geikie, Dana, Gresley, Grimsley, and 
Chadwick. Others have regarded the structure as being 
due to crystallization or to "imperfect crystallization." 
Owen, Newberry, Geikie, Sach, Grabau, and Keyes have 
advocated one or both of these methods. 

Odd suggestions have been made by Sorby, who sup- 
posed that oolites had formed the sides of the cones ; by 
Daintree, who believed the cones were chemical precip- 
itates ; and by Young, who says unequivocally, that they 

1 The writer -wishes to acknowledge his indebtedness to many friends for 
material furnished for this study. Mr. H. L. Griley and W. H. Twenhofel 
have furnished some excellent material. The late G. C. Broadhead collection 
at the University of Missouri contained several excellent small specimens. 

200 W. A. Tarr — Cone-in-Cone. 

are due to gases rising through the muds. A view similar 
to that of Young has been expressed by Lawson. This 
view has been adequately disproved by Gresley and 

After a careful review of the literature on cone-in-cone 
the writer feels that there are still many points upon 
which more data are desired. He has studied a suite of 
specimens from many places, and it is the object of this 
paper to give the results attained, in the hope that others 
will give some attention to a very difficult, yet extremely 
interesting problem. 

Description of Cone-in-Cone. 

The general features of cone-in-cone are essentially the 
same for all occurrences. This similarity is true even of 
the minor features of the structure. The structure con- 

Fig. 1. Fig. 2. 

Fig. 1. — Cone within cone. 
Fig. 2. — Single cone. 

sists of a series of right circular cones, which may fit 
one inside the other (fig. 1). This is the common mode 
of occurrence, but cones may also occur singly (fig. 2). 
They are usually grouped along a plane (fig. 3), which, 
if it occurs on the surface of a concretion, may be curved 
to fit the surface (fig. 4). Cone-in-cones are usually asso- 
ciated with concretions and in areal extent are generally 
co-extensive with them. The writer has observed this 
structure in concretions five feet in diameter. The cones 
are generally perpendicular in the central part of the 
layer (at a in fig. 4), but near the edge they are inclined 
(as at b, fi.g. 4). 

W. A. Tarr — Cone-in-Cone. 


The height of the cones varies from a thirty-second of 
an inch to eight or nine inches. Those from one to four 
inches in height are the most common. The diameter of 
the base depends upon the height and upon the angle of 
slope of the sides of the cones. In many cones this 
diameter is nearly equal to the height of the cone. The 
included angle (a, fig. 2) at the apex of the cone ranges 
from 25° to 70°. If the cone is well developed this angle 
is generally 60° to 70°. Partially developed cones are 

Fig. 3. 

Fig. 3. — Cones along a plane. 

The sides of the cones are rarely perfectly smooth. 
Striatums such as are seen on slickensided surfaces are 
common. What Gresley calls " conic scales" (fig. 5) are 
also very common. The sides of the cone may be ribbed 
or fluted, thus giving a notched outline on the base (fig. 6). 


Fig. -i. — Concretion with layer of cone-in-cone (scale 1 inch equals 1 foot). 

The inside of a cone into which another cone fits is 
always ribbed with circular rings, which are darker in 
color than the material composing the mass of the cone- 
in-cone (fig. 7). These rings are always on the inside of 
the cone. They vary in width from mere lines to over 
one quarter of an inch. Within a given cone they are 
very fine near the apex, and coarsest near the lower edge. 
This holds true for all specimens examined by the writer 


W. A. Tarr — Cone-in-Cone. 

and has been so described by others. These rings are 
composed of clay, usually dark as mentioned above, and 
are free from carbonates. The ring of clay occupies a 
depression on the inside of the cone (fig. 8, A and B). 
This clay is similar to the insoluble residue of the cone- 
in-cone. The rings are broadest and most numerous in 
the more impure specimens of cone-in-cone. The ridges 
between the rings are striated and have the same slope 
as the outer surface of the cone, fitting into the cone-cup. 

Fig. 5. 

Fig. 6. 

come sea 

Fig. 5. — Conic scales on the side of cone. 

Fig. 6. — Basal outline of a cone, with conic scales on the sides. 

Fig. 7. 

Fig. 7. — Rings in cone-in -cone. 

The cones may be oblique especially near the edge of 
a cone layer. Some may have the apex removed and 
others may be more or less bent and twisted. A peculiar 
feature of the cones in layers, not associated with con- 
cretions, is the flaring of the base and the acute apex 
(see fig. 9) . Such cones are common in Bond Co., Illinois, 

W. A. Tarr — Cone-in-Cone. 


and in some thin lenses of cone-in-cone in Boone Co., 
Missouri. The fibers composing the cone-in-cone layers 
are parallel or inclined. If the latter, they are parallel 
to the surface of the cones. 

Fig. 8. 

Fig. 8. — Section of clay rings showing relationship to fibers of calcite. 
A — natural size. B — enlarged three times. 

The composition of cone-in-cone is significant. -Anal- 
yses show from 60 to 98 per cent CaC0 3 , with the 
remainder usually clay and other insoluble materials. 
The common occurrence of cone-in-cone in shales and 

Fig. 9. 

Fig. 9. — Cone showing flaring of base. 

always in association with calcareous portions of the 
shales would account for the presence of the clay. Prob- 
ably the association of cone-in-cone with calcareous con- 
cretions has a bearing on their occurrence. The writer 
has found them in thin lenses of fibrous calcite in the 
Pennsylvania!! shales in Boone Co., Missouri. 

204 W. A. Tarr- — Cone-in-Cone. 

A point of much value in discussing the origin is 
whether the cone-in-cone is dominantly calcite or aragon- 
ite. Where determinable in the writer's specimens the 
material is calcite. Cone-in-cone occurs with ferruginous 
concretions in various localities, hence it is possible that 
some cone-in-cone is ferruginous, but no cone-in-cone com- 
posed of siderite has been seen by the writer or reported 
in the literature. 

Mode of Occurrence of Cone-in-Cone. 

Cone-in-cone usually occurs in shales in association 
with concretions. It may occur as a layer on the upper 
side of the concretion, as bands within them, or as a layer 
on the lower side. The bands within the concretions 
usually extend from one side to the other, and may be 
several inches wide. There may be several bands in one 
concretion. The layers of cone-in-cone above and below 
a concretion may attain a thickness of several inches, but 
they usually thin out at the edges of the concretion. 
Young claims that the cone-in-cone may fold back upon 
itself, thus forming two bands of cone-in-cone with their 
apexes pointing toward each other. Such parallel bands 
are fairly common, but there is no evidence to support 
Young's view as to the cause of the parallelism. 

The structure may occur as one or more layers, not in 
association with concretions. The occurrences in the 
Devonian shales in Pennsylvania, described by Gresley, 
show three to five layers associated with a band of sand- 
stone. Such layers may have great variability in plan 
as well as diverse shapes when seen in section. 

More rarely single layers of calcite showing cone-in- 
cone may occur within shales. The layers range from 
one half inch to two and one half inches in thickness. 
The cone-in-cone may extend entirely across the layer or 
it may be developed on one side only. 

A very rare occurrence is in coal. Well developed 
cones from four to six inches high are found in the British 
coal beds. Their sides are highly polished and slicken- 

W. A. Tarr—Cone-in-Cone. 205 

Distribution of Cone-in-Cone. 

Geo graphic ally cone-in-cone structure has been 
reported in various places in the United States. The 
localities first noted were those in western New York, 
Pennsylvania, and Ohio. Other localities are in Michi- 
gan, Illinois, Iowa, Missouri, Texas, Kansas, Nebraska, 
South Dakota, Wyoming, Montana, Colorado, Utah, and 
California. (The writer's collection includes specimens 
from the states in italics.) It is probable that the struc- 
ture is more widely distributed than is indicated above, 
for its occurrence is only mentioned incidentally in most 
of these cases and there is usually no reference to it in 
the index of the reports. It is known to be rather widely 
distributed in England, France, and Germany. 

Geologically there is apparently nothing significant 
about the geological occurrence of cone-in-cone structure. 
It has been reported in middle Cambrian beds in Utah. 
It occurs in the Devonian in New York, Pennsylvania, 
Ohio, and Michigan. The Pennsylvanian in Illinois, 
Iowa, Missouri, and Kansas has furnished numerous 
specimens of cone-in-cone. The Permian in Kansas, 
Montana, and South Dakota, and the Eocene in Texas 
have been found to contain the structure. In Europe it 
has been found principally in the Carboniferous and the 
Jurassic formations. 


The published views regarding the origin of cone-in- 
cone are extremely vague. Pressure is regarded by some 
as a factor, but why pressure should produce cone-in-cone 
is not explained. The presumption is that pressure in 
forcing the material of the structure through the adjacent 
rock produced cones. Crystallization is also suggested, 
but why it should develop cones is again not explained. 
Many minerals are known which may crystallize in radiat- 
ing masses, but such masses are spherical, or closely 
related forms, and not conical. It is not clear- why 
crystallization should develop only a fraction of a sphere 
and do it as perfectly as has been done in the case of 

The following suggestions represent the conclusions to 

206 W. A. Tarr — 

which the writer has come from his past and recent 
studies. Future studies will call for much revision and 
possibly the discarding of parts or all of these sugges- 

The suggestion made here is that cone-in-cone structure 
has been formed through a combination of pressure, more 
or less localized, and solution. Contributing factors are 
both the radial and parallel arrangement of the crystals 
of calcite composing the structure, and the cleavage of 
the original carbonate. 

Evidence of movement — The fibrous calcite constituting 
a layer of cone-in-cone is often banded horizontally. 
When the structure is developed in such a layer these 
bands are displaced by the cone-in-cone. This displace- 
ment is analogous to a fault, but careful study showed 
that the layer has not been faulted along a plane ; it is 
the cone which has moved. Another proof of such move- 
ment is the occurrence of the cones within each other. 
A large cone (fig, 10) may contain a small one which 
has penetrated the large one, yet has not displaced the 

Fig. 10. 

Fig. 10 — Shows small cone penetrating large cone which does not show 
any evidence of displacement in the large cone. 

outer surface of the large cone. The inward movement 
may amount to from one quarter to one half an inch in 
a cone which is two to two and one half inches high. 

Evidence of solvent action — The last statement in 
regard to movement also furnishes proof that some 
material has been removed through solvent action. The 
small inner cone could not have moved inward without 
displacing the material of the larger cone. As there is 
no external evidence of such displacement the cone must 

W. A. Tarr—Cone-in-Cone. 207 

have been enabled to move because of the solution of the 
material previously occupying the space between the 
small cone and the large one. This would be the material 
in the shaded area in figure 11. The removal of this 
material m solution would enable the cone to penetrate 
the larger cone. 

Fig. 11. 

Fig. 11. — Shaded area shows the material removed in solution, to permit 
the movement of the inner cone. 

The rings of dark clay on the interior of the cone cup 
have lost all their carbonate content and represent merely 
the insoluble residue obtained by dissolving the material 
composing the cone. A careful study of these rings 
showed that they contained more material near the base 
of the cone, due to the fact that more carbonate material 
has been removed by solution from that portion of the 
cone. The ring of dark clay occupied a depression on the 
inside of the cone cup. These depressions may be one- 
eighth inch or more deep, and are separated from each 
other by a thin partition of calcite (fig. 8, B). The ring 
of clay fits into this depression. 

The facts cited above prove that movement and solu- 
tion have taken place during the development of the cone- 
in-cone structure. The conclusion may be drawn that 
the movement certainly involved some pressure, and as 
material was removed by solution the cone slipped down 
(or up) inside the larger cone. 

Source and amount of the pressure — It can hardly be 
doubted that some pressure was involved in the formation 
of the structure. This pressure probably was not large, 
because most occurrences of cone-in-cone were never 
deeply buried. Many have thought that the pressure, 
where the cone-in-cone was associated with concretions, 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 21. — September, 1922 


208 W. A. Tarr — Cone-in-Cone, 

was due to the growth of the concretion. If the concre- 
tion actually developed in place a certain amount of pres- 
sure would have developed. However, the writer 1 has 
proved that certain concretions occurring with cone-in- 
cone over them were syngenetic in origin, hence no pres- 
sure would have developed in such cases. If it can be 
assumed that the growing strength of a calcareous con- 
cretion is equal to the crushing strength of similar mate- 
rial, the pressure so produced would be considerable, but 
would be dependent upon the compressibility of the 
enclosing materials. That some pressure would be pro- 
duced in the case of ' epigenetic concretions cannot be 
denied, but some other source of pressure must be sought 
when the cone-in-cone is not associated with concretions, 
or when the concretions are syngenetic. The cone-in- 
cone unassociated with concretions is as perfectly devel- 
oped as when associated with them. This is especially 
true of cone-in-cone in coal. 

Another possible source of pressure exists, if the 
original form of the calcium carbonate was aragonite and 
had been changed to calcite. This change involves a 
volume increase of 8.35 per cent, sufficient to produce 
considerable pressure. But is there any reason for 
assuming that the original form was aragonite! No 
cone-in-cone the writer has examined is composed of 
aragonite now. This very fact, however, may mean that 
cone-in-cone develops only where aragonite has altered 
to calcite as will be discussed below. All cone-in-cone 
structures show fibrous calcite, but that this was origin- 
ally aragonite cannot be proved. Fibrous calcite from 
Boone Co., Missouri has been sold to mineral dealers as 
aragonite, showing that this structure was regarded as 
evidence of the material being aragonite. The develop- 
ment of cone-in-cone only in a crystalline, fibrous calcite 
is significant, however, and certainly suggests that arago- 
nite . was the original . form of the calcium carbonate. 
According to Clarke 2 " aragonite alters easily to para- 
morphs of calcite, but the reverse change rarely, if ever, 
occurs/' Calcite is the more stable form. 

It does not seem likely that either the growth of concre- 
tions or the weight of overlying sediments would be an 

1 Tarr, W. A. : Syngenetic Origin of Concretions in Shales, Bull. Geol. 
Soc. Am., vol. 32, pp. 373-384, 1921. 

2 Clarke, F. W., Data of Geochemistry, U. S. G. S., Bull. 695, p. 549. 

W. A. Tarr—Cone-in-Cone, 2G9 

important source of pressure. The change of aragonite 
to calcite would give an adequate source of pressure and 
seems to be the most likely source. What sources of 
pressure other than this exist in the rocks associated with 
the cone-in-cone are not evident to the writer. Those 
which are there certainly function, as it is doubtful if 
much movement could have taken place without some 

Origin of the Conical Form. 

The origin of the conical shape of the cone-in-cone is 
extremely puzzling. If the angle of the slope of the cones 
was always the same, and if it was similar to the cleavage 
angle of the material composing it, the problem would be 
much simpler. But the cleavage angle in calcite is 
approximately 105°, which is far greater than that of 
the cone-in-cone. This may mean that the cone-in-cone is 
not related to the cleavage, or, that the original substance 
was not calcite and thus had a different cleavage angle. 
As the original substance was calcium carbonate, it may 
be that aragonite was the original mineral. But the best 
cleavage in aragonite is pinacoidal whereas the cleavage 
parallel to the brachydome is poor. Since a series of 
brachydomes is possible on aragonite, forms with acute 
domes might have a poor cleavage with angles similar to 
those of the cone-in-cone. The poor cleavage of aragonite 
along this plane, however, does not favor the possibility 
that the initial form of the cone-in-cone was due to it. 

Another possibility, if the original form of the material 
was aragonite, should be considered. Radiating struc- 
tures are very common in aragonite. This structure may 
give rise to stalactites, spherical masses, or radiating 
tufts. The tufts have considerable significance in this 
problem. If the mineral were deposited along a plane 
with deposition occurring at many points, a layer con- 
sisting of an aggregate of such radiating tufts would 
result. The spacing of the tufts along the plane would 
determine their size, and incidentally the angle of the 
outermost fibers on the tuft. Close spacing would give a 
sharp angle, and wide spacing a low angle. If, now, 
pressure were brought to bear on the layer, movement 
would be greatly aided by the fact that the outermost 
layers would have the distinct cleavage of the pinacoidal 
and prismatic faces. The fact that the cone surface is 

210 W. A. Tarr — C one-in-C one. 

parallel to these radiating fibers is compatible with this 
suggestion. Giimbel has made an interesting suggestion 
along these lines. He thinks that water might find its 
way through a bed and form stalactites on the under side, 
each stalactite having a cone shape. 

It should be noted that the most perfect cones are 
always developed in radiating aggregates of fibers. 
When the fibers are parallel the peculiar cone with a 
flaring base and a long slender body develops. 

It is very probable that those who have expressed the 
view that cone-in-cone was due to crystallization were 
thinking of radial aggregates as the initial cause of the 
conical form. Such radial aggregates may be seen in 
concretions. And it cannot be said that crystallization 
might not form aggregates which would be the real cause 
of the cone-in-cone structure. Such a radial aggregate 
would provide the form, without resorting to cleavage as 
a factor in initiating movement in the cone-in-cone. 

Aside from the possibility of the influence of radial 
fibrous aggregates and the cleavage of aragonite, both of 
which would influence the initiation of the cone-in-cone 
while the material was still aragonite, there are two other 
suggestions as to the origin of the conical form that are 
of value. One is dependent upon pressure and the other 
upon the alteration of the aragonite to calcite. 

That pressure alone might produce cones was suggested 
to the writer as the result of testing some cubes of lime- 
stone to determine its crushing strength. Perfect and 
imperfect cones were formed during the tests, the largest 
ones being two inches across the base. Many were 
remarkably perfect and showed several features seen in 
the cone-in-cone. The included angle at the apex of the 
cone was from 30° to 60° and in some cones the angle 
was greater near the base than at the apex, a feature 
noted in some cone-in-cone. In the mechanical produc- 
tion of cones there also may be seen an explanation of 
cone-in-cone structure, because, on the base of the larger 
cones mentioned above, there were numerous minute cir- 
cular cracks, which indicated that another cone might 
develop within the first. When the fibers of the calcite 
are radial, pressure upon the base of such a group might 
easily induce a fracture, more or less parallel to the fibers, 
which would be circular. Thus the outline of the cone 
would be inaugurated. Solution along this fracture 

W. A. Tarr—Cone-in-Cone. 211 

would then become a very important factor. The cones 
formed in coal were certainly due to pressure unaided 
by solution. 

The other method by which a conical shape might be 
induced hinges upon the original form of the calcium 
carbonate being aragonite and subsequently altering to 
calcite. Conceivably this alteration, being brought about 
by water, would start along the joint in a bed of cone-in- 
cone and proceed inward. As the aragonite changed to 
calcite there would be a volume increase and expansion 
of the outer portion, leaving a suggestion of a cone on 
the interior. How effective this would be is problemati- 
cal ; very likely the effect would be slight. 

But a combination of the above factors could very well 
give rise to the cone-in-cone structure. The radial aggre- 
gates, when acted upon by pressure, would favor the 
development of fractures along which solutions might 
enter, dissolve, and remove material, thus permitting 
further movement. Artificially formed cones, taken in 
conjunction with the fact that a majority of cone-in-cone 
occurs in layers of radial fibers, lends strong plausibility 
to this suggestion. 

The development of cone within cone might conceivably 
be accomplished in the following manner. Solutions 
entering along a joint might convert the aragonite to 
calcite. This change would progress inward from the 
joint a variable distance, where another fracture plane 
would develop by the expansion of the material in chang- 
ing from aragonite to calcite. Thus a series of cones 
might be formed, one within the other. Another method 
has been indicated above, that is, pressure producing a 
series of minute cracks on the base of a cone. These 
were parallel to the circumference of the base and con- 
tinued pressure might easily have started another cone 
within the first. But the most probable method would 
seem to be by means of the radial structure of the calcite. 

Was the Material Originally Aragonite? 

Reference has been made so frequently in the above, 
discussion to the possibility of the calcite having been 
originally aragonite that it would seem well to summarize 
the reasons for believing that this may have been the case. 

212 W. A. Tarr — Cone-in-Cone, 

1. The cone-in-cone is always found in material of fibrous 

form, a structure more common to aragonite than calcite. 

2. The fibers are dominantly radiating, also a common 

arrangement for aragonite. 

3. Aragonite easily alters to the paramorph calcite. 

4. The change from aragonite to calcite involves a volume 

increase of 8.35 per cent, a change that would give a 
marked pressure, and pressure has been involved in the 
formation of the cones. 

5. Radial aggregates with pinacoidal cleavage (both common 

in aragonite) would be favorable factors in the formation 
of the cones. 

6. Aragonite is more readily soluble in groundwaters than 


7. The combination of factors given above would hardly be 



Cone-in-cone is best developed in shales and is gener- 
ally composed of calcinm carbonate in the form of calcite, 
it having been originally, possibly, aragonite. The cal- 
cium carbonate always contains more or less argillaceous 
material. Studies show that there must have been some 
movement in the development of the cone structure and 
also that a certain amount of solvent work had taken 
place. The insoluble material in the calcite is left behind, 
through the removal of the calcium carbonate, and forms 
the rings of material (usually dark) which are seen on 
the interior of all cones that are well developed. The 
fibrous structure of the calcium carbonate has favored the 
parallel arrangement of the rings within the cone. 

The cones are probably the result of pressure due to 
(1) weight of overlying sediments, (2) growth of concre- 
tions, (3) (the most important) an increase in volume due 
to the change of aragonite to calcite, acting in connection 
with the radial or parallel arrangement of the calcium 
carbonate fibers ; and to solution, which became effective 
when the pressure had induced the first crack in the layer, 
groundwater entering and removing the material along 
this joint. The removal of material along this fracture 
permitted the further movement of cone within cone. 

Mineralogical Laboratory, 
University of Missouri, 
Columbia, Mo. 

W. A. Tarr— Cone-in-Cone. 213 

Bibliography of Cone-in-Cone. 

Barbour, Carrie A. : Neb. Aead. Sc. Proc, vol. 7, pp. 36-38, 1897. 

Broadhead, G. C. : Cone-in-Cone, Science, vol. 26, p. 15, 1919. 

Chadwick, G. B. : Bull. Geol. Soc. Am., vol. 32, p. 26, 1921. 

Daintree, E.: Quart. Jour. Geol. Soc, vol. 28, p. 283, 1872. 

Dawson, J. W. : Acadian Geol., pp. 676-677, 1868. 

Garwood, E. J. : Geol. Mag., vol. 9, p. 334, 1892. 

Geikie, Sir A. : Textbook of Geology, 4th edit., pp. 421, 1902. 

Grabau, A. W. : Prin. of Stratigraphy, pp. 788-789. 

Gresley, W. S.: Geol. Mag., vol. 9, p. 432, 1892. 

Gresley, W. S. : Geol. Mag., vol. 4, p. 17, 1887. 

Greslev, W. S. : Cone-in-Cone, Quart. Jour. Geol. Soc. London, vol. 50, pp. 

731-739, 1894. 
Grimsley, G. P.: Mich. Geol. Sur., vol. 9, pp. 100, 109, 1903-1904. 
Harker, Alfred: Geol. Mag., vol. 9, p. 240, 1892. 
Hildreth, S. P. : this Journal, vol. 29, pp. 99-100, 1836, plate 14. 
Jukes: Manual of Geology, 1872. 

Keyes, C. K.: Cone-in-Cone, Proc. Iowa Acad. Sc, vol. 3, pp. 75 L 76, 189. 
Lawson, A. C. : Bull. Geol. Soc Am., vol. 32. 
Leonhard, K. C. von: Charakteristik der Felsarten, p. 418, 1823. 
Marsh, O. C. : Proc. Am. Assoc Adv. Soc, vol. 1, p. 211, 1873. 
Newberry, J. S. : Geology of Ohio, vol. 1, p. 211, 1873. 
Owen, D. D. : Eeport of the Geol. of Wis., Iowa, and Minn., p. 112, 1852. 
Sach, A. J. : Geol. Mag., vol. 9, p. 505, 1892. 
Sorby, H. C. : "On the Origin of Cone-in-Cone, ' ' Brit. Assoc. Adv. Sc, vol. 

29, p. 124, 1859. 
Young, John: Geol. Mag., vol. 9, pp. 138, 278, 480, 1892. 

214: R. W. Chaney — Flora of Payette Formation. 

Art. XXL — Notes on the Flora of the Payette For- 
mation; by Ralph W. Chaney, Carnegie Institution 
of Washington. 

The Payette flora as described by Knowlton in 1898 1 
is made up of 32 species. It was referred by him to the 
Miocene because of its resemblance to the Upper Clarno 
and Mascall floras of the John Day Basin which were 
supposed at that time both to be of Miocene age. Later 2 
the same author determined the age of the Payette as 
Upper Eocene, because of its resemblance to the Upper 
Clarno flora which his later "work had shown to be of 
Upper Eocene age. 

The present paper is based on observations and collec- 
tions made in southwestern Idaho and adjacent Oregon 
during the season of 1921 under the auspices of the Car- 
negie Institution of Washington. Its purpose is to add 
to the number of forms of fossil plants known from the 
horizon, to present further evidence regarding their age, 
and to make suggestions regarding the conditions under 
which they lived. 

The Payette formation comprises several hundred 
feet of sediments overlaying a basalt series which 
appears to be an eastward continuation of the Colombia 
lavas. According to Buwalda 3 its basal portion is inter- 
stratified with these lavas on the south side of the 
Snake river valley. Where studied in connection with 
its plant fossils, the Payette is made up largely of a 
fine gray salmon-colored shale, highly indurated. In the 
vicinity of Montour, Idaho (the Marsh postoffice locality 
of Knowlton and Lindgren), there are conspicuous white' 
shales made up of diatomaceous sandy material. Lig- 
nite seams are of common occurrence, and locally the 
sediments appear to be made up of broken-down granite. 
Fossil plants are found most frequently on small slabs 
of the highly indurated shale which have been weathered 
loose and scattered over the slopes of the hills. They 
have been collected by the writer at the following 
localities : — 

1 Knowlton, F. H. ; Fossil Plants of the Payette Formation, U. S. Geol. 
Survey, 18th Ann. Kept., Pt. 3, pp. 721-744, 1898. 

2 Knowlton, F. H. : Fossil Flora of the John Day Basin, Ore., U. S. Geol. 
Survey, Bull. 204, pp. 110-111, 1902. 

3 Buwalda, John P.: Letter of Nov. 6, 1921. 

R. W. Chaney — Flora of Payette Formation. 215 

1. 12 miles west of "Weiser, Idaho on the hills north 
of the Snake river. 

2. Ballantyne ranch, % of a mile southeast of the 
ranch house in a group of white hills. The ranch is 13 
miles southwest of Homedale, Idaho, and this portion of 
it is in Malheur Co., Oregon. 

3. Ballantyne ranch, along the ridge running north- 
west of the ranch house. 

4. Sucker creek, 5 miles southwest of The Rocks, 
Idaho, on the hills south of the valley, near the Idaho- 
Oregon line. 

5. Sucker creek, 4% miles southwest of The Rocks, 
and on the north side of the valley. 

6. Cartright ranch on Shafer creek, 7 miles south of 
Horseshoe Bend, Idaho, in a gulch about y 2 mile south- 
west of the ranch house. 

The locality west of Weiser is about 75 miles north of 
the southernmost locality on Sucker creek, and is 50 miles 
west of the Cartright ranch which lies farthest east. 

Of the material collected, some fifteen forms are 
reserved for later treatment because of their fragmentary 
nature. Of the material which is readily determinable, 
there are 30 species which may be tabulated as follows : — 

Number of species included in previously described 

Payette flora 13 

Number of new species 4 

Number of species previously known and referred 
for the first time to Payette 13 

Total 30 

It will be seen that the previously described flora is 
increased by 17 species, 4 of which are new, the whole 
representing an addition of over 50 percent. That only 
13 species or less than half of those previously known 
from the Payette are represented in my collections may 
be explained by the fact that only one of the localities 
from which Knowlton secured his material was visited, 
and this one was represented by only four species in 
his collections. The total number of forms known from 
the Payette may therefore be placed at 49, of which 17 
are here recorded for the first time : — 

216 R. W. Chaney — Flora of Payette Formation. 

Acer gig as 
Acer n. sp. 
Aesculus simulata 
Odostemon simplex 
Betula n. sp. 
Glyptosir obits europceus 
Laurus princeps 
Finns knowltoni 

Platanus dissecta 
Prunus n. sp. 
Quercus clarnensis 
Quercus day ana 
Quercus n. sp. 
Salix californica 
Salix perplexa 
Sapindus oregonianus 

Vlmus calif ornica 

Regarding the four new species, it has been thought 
best to delay describing and figuring them until further 
collections have been made and their relations are more 
fully understood. In so far as they resemble fossil or 
living forms , they will be considered in correlation, and 
their ecological significance also will be discussed below. 

The following table shows the distribution of the 21 
species which are common to the Payette and to other 
localities where the age of the beds is known with 
reasonable certainty : — 







1 - 

































Acer gigas 


Aesculus simulata 


Glyptostrobus europaeus 


Juglans oregoniana 


Laurus princeps 


Odostemon simplex 


Pinus knowltoni 


Platanus dissecta 






Populus lindgreni 



Quercus clarnensis 


Quercus consimilis 


Quercus dayana 



Quercus simplex 



Quercus simulata 


Salix angusta 



Salix californica 


Salix perplexa 



Sapindus or-egonianus 


Sequoia angustifolia 



Ulmus californica 





Ulmus speciosa 



Total species 











R. W. Chanei/ — Flora of Payette Formation. 217 

As in the case with most of the Great Basin floras, the 
Payette flora most closely resembles the floras from the 
Upper Clarno and Mascall horizons of the John Day 
Basin, with 5 species and 12 species in common, respect- 
ively. This is dne both to the close geographic relation 
of the two localities and to the fact that the floras from 
the John Day Basin have been more fnlly described 
than from any other Great Basin locality. In consider- 
ing the age of the species common to the Payette and 
other floras, the following summary makes the relations 
clear : — 

Payette species occurring in the Eocene 8 

Payette species occurring in the Oligocene 5 

Payette species occurring in the Miocene 14 

Of the 8 Eocene species, 4 are restricted to the Eocene, 
while 1 is found also in the Oligocene and 3 in the 
Miocene. Of the 5 Oligocene species, 2 are restricted 
to the Oligocene, while 1 is found also in the Eocene and 
2 in the Miocene. Of the 14 Miocene species, 9 are 
restricted to that horizon, while 3 occur also in the Eocene 
and 2 in the Oligocene. A summary of the number of 
Payette species which are restricted to the several 
Tertiary systems shows : — 

Payette species found elsewhere only in the Eocene 4 

Payette species found elsewhere only in the Oligocene .... 2 
Payette species found elsewhere only in the Miocene 9 

Several other aspects of the stratigraphic relations 
of the Payette flora should be considered. Of the forms 
doubtfully determined and not here listed, Ilex cf. leonis, 
and Celastrus confluens? are Miocene species, while 
Rhus cf. my riccE folia and Betula heteromorphaf are typi- 
cally Eocene. Of the new species, a species of Prunus 
may be related to two species of the same genus from the 
Mascall, and species of Betula, Acer, and Quercus find 
their closest relation to living species. Of still greater 
importance is the bearing on the problem of a flora 
recently collected at Austin, Oregon, which appears to 
be certainly from the Mascall formation of Miocene age. 
In this flora, associated with typical Mascall species, 
are Quercus consimilis, Quercus clarnensis, and Ulmus 
speciosa, all of which are typical Upper Clarno species. 

218 R. W. Chaney — Flora of Payette Formation. 

If these three species are thus shown to range into the 
Miocene, the preceding summaries must be changed, as 
follows : — 

Payette species occurring in the Eocene 8 

Payette species occurring in the Miocene 17 


Payette species found elsewhere only in the Eocene 2 

Payette species found elsewhere only in the Miocene .... 9 

• From a numerical standpoint, the Payette flora is more 
closely related to the Miocene floras than to those of 
any other Tertiary system. However, mere numbers of 
common species are not, in themselves, the sole criteria 
for placing the age of a flora, The two species which 
remain to make up those limited to the Eocene are Odos- 
temon simplex and Quercus simplex. The former is 
represented in the collections from the Upper Clarno 
at Bridge Creek by a single specimen; it is closely similar 
to several of the living barberries. Quercus simplex 
differs only slightly from several other species of west- 
ern fossil oaks which range up into the Miocene. 
Neither of these species carries great weight in indicating 
Eocene age. Again both Quercus simplex and Salix 
angusta are species with somewhat indefinite character- 
istics, so that the record of the range of the former from 
the Raton to the Upper Clarno and of the latter from the 
Green River to the Mascall may be in each case in error. 
On the other hand, the presence in the Payette of species 
which are characteristic of several localities of the Mio- 
cene gives more weight to an age reference to that hori- 
zon than the occurrence in it of several species from 
another horizon which are of restricted distribution and 
rare occurrence. In this regard, Platanus dissecta and 
Ulmus calif omica, both originally described from the 
auriferous gravels, 4 are typical Miocene species. By 
reference to the tables of distribution, it will be seen 
that Platanus dissecta has been found in the Miocene 
of the John Day Basin (Mascall formation), of Table 
Mountain, California (Auriferous Gravels), of Corral 
Hollow, and of central Washington (Ellensburg forma- 

4 Lesquereux, Leo : Fossil Plants of the Auriferous Gravel Deposits of 
the Sierra Nevada, Mus. Comp. Zool. Mem., vol. 6, No. 2, pp. 1-62, 1878. 

R. W. Chanei/— Flora of Payette Formation, 219 

tion). Although it also occurs in the Chalk Bluffs 
locality of the Auriferous Gravels, which is commonly 
considered to be of Eocene age, Lesquereux states (4, p. 
11) that it is more characteristic at the Table Mountain 
locality. A recent study of the type material of Acer 
bendirei, one of the most typical Miocene species in the 
west, tends to show that it is to be referred to the genus 
Platanus rather than Acer, and that certain specimens 
of Platanus dissect a are identical with it. Platanus 
dissect a is one of the more common species of the Pay- 
ette flora, and is of wide occurrence in the Miocene of 
several adjacent states. Clearly it is a strong indicator 
of the Miocene age of the Payette. Likewise, Uhnus 
calif o mica of the Payette flora has a Miocene distributioD 
in the John Day Basin, at Table Mountain, California, 
and at Ellensburg, Washington. It, too, occurs in the 
Eocene horizon of the Auriferous Gravels, probably as 
a geologic pioneer, for it is not recorded elsewhere from 
the Eocene. In the same way, Sapindus oregonianus, 
which is one of the more abundant species of the Payette, 
is found in at least three widely separated Miocene 
localities of Oregon, and has never been noted outside 
of this system. 

A close similarity between certain Payette species 
and living forms is again taken as an indication of the 
closer relation of the Payette flora to the Miocene 
than to the Eocene. The similarity of three new species 
to modern representatives of their genera has already 
been mentioned; Quercus n. sp. is strikingly like the 
western Q. clirysolepis now living in California; Betula 
n. sp. closely resembles B. occidentalis of the Pacific 
Coast ; and Acer n. sp. is hardly to be distinguished from 
Acer glabrnm, also a tree of the west. Of the previously 
described species. Knowlton comments 5 on the simi- 
larity of Aesculus simulata to Ac. oclandra and glabra 
of the eastern United States ; he names Populus eotremu- 
loides* having in mind a close relation with the living 
P. tremuloides; certain specimens of Populus lindgreni 
have a striking resemblance to living poplars of 
the balsamifera type; Lesquereux 7 mentions a simi- 

5 Knowlton, F. H. : Fossil Flora of the John Dav Basin, Ore., p. 78. 

6 Knowlton, F. H. : Fossil Plants of the Payette Formation, p. 725. 

' Lesquereux, Leo : Fossil Plants of the Auriferous Gravel Deposits, pp. 
11 and 16. l ' F1 

220 R. W. Chaney — Flora of Payette Formation. 

larity between Plat anus dissecta and P. occidentalis now 
living in the eastern half of the United States, and 
between TJlmus calif or nica and U. alata of the south- 
eastern United States. In the case of Plat anus dissecta, 
its resemblance to the western species, P. racemosa, seems 
even closer than to P. occidentalis. This resemblance 
between various forms of the Payette flora and living 
plants adds further weight to the reference of the Payette 
to the Miocene rather than to an older period. 

A probable explanation of the conflicting evidence of 
age as shown by Eocene as well as Miocene representa- 
tives in the Payette flora is that the disparity in age 
is more apparent than real. The writer has not yet 
brought together all the necessary evidence, but recent 
field studies in the John Day Basin and at other points 
in eastern and central Oregon point to the placing of 
the Upper Clarno formation in the Oligocene rather than 
the Eocene. If this is correct, and the supporting evi- 
dence will be brought forth in a subsequent paper, there 
are no typical Eocene plants in the Payette flora, but 
rather an increased number of Oligocene forms which 
may be supposed to have been precursors of the Miocene 
flora to follow. 

The position of the Payette formation, above basalt 
lavas, is in accord with the relation of the Miocene Mas- 
call formation to the Columbia lavas in Oregon. Ver- 
tebrate remains collected from the Payette formation 
by Buwalda are referred by him to the Middle or 
Upper Miocene 8 . There appears, therefore, to be an 
agreement between the evidence of the flora, the fauna, 
and the stratigraphy in pointing toward the Miocene 
age of the Payette formation. So far as the plants are 
concerned, the age would not be younger than Middle 
Miocene, and in view of the inclusion of Oligocene and 
possibly Eocene forms, it might better be considered 
as Lower Miocene. The present knowledge of the west- 
ern floras does not justify an' attempt to draw the lines 
of age too sharply, and at this time the writer is satis- 
fied to make the reference to the Miocene without further 

The Payette flora has a composition so like that 
of the forests in certain parts of the Pacific Coast region 

8 Buwalda, John P. Letter of Feb. 10, 1922. 

R. W. Chanei/— Flora of Payette Formation. 221 

today that the physical conditions under which it existed 
may be postulated with a reasonable degree of accuracy. 
The most dominant genus is Quercus, with 7 species. 
It may be said, although further mention of the taxo- 
nomic aspects of the flora is not within the province of 
this paper, that several of these species can hardly be 
considered distinct in view of the large number of inter- 
grading specimens and the great variation among living 
oaks of a single species. But the genus Quercus is 
dominant even without regard to the number of species 
in that its individual leaves outnumber all others together. 
There is a close resemblance of all the fossil species of 
oak to the type represented by the living species Quercus 
cliry sole pis and Q. sadleriana. Both of these species 
occupy the dry slopes of the upper foothills and summits 
of the coast ranges and the west slopes of the Sierras 
and Siskiyous of northern California and southwestern 
Oregon. Their fossil relatives appear to have had the 
coriaceous texture which is characteristic of leaves in 
such a habitat, and to have made up a forest on the slopes 
of an area of high relief and consequent exposure. And 
while these oaks may have ranged down into the canyons, 
as does Quercus chrysolepis today, the typical stream 
border flora comprised Platanus dissecta, Acer n. sp. 
Juglans oregoniana, and the various species of Salix 
which are found in the Payette. Swampy stream or 
lake borders are indicated by the fragments of Typha 
(?), grasses, and Equisetum. The local dominance of 
Populus eotremuloides, which is found with birches of the 
occidentalis type only at the Cartright ranch locality, may 
be due to the presence of certain moist slopes, or possibly 
to a forest fire and the subsequent development of a burn 
subsere. Sequoia angustijolia, which is abundantly 
represented in the Mascall flora of the John Day Basin, 
is of rare occurrence in the Payette, suggesting a lower 
humidity. In accord with the idea of a dry slope habitat, 
as suggested by the abundance of chrysolepis oaks and 
the scarcity of Sequoia, is the occurrence of Pinus knowl- 
toni and the scrub species of Odostemon, which is closely 
related to the living Oregon grape, and of Castanopsis, 
Philadelphus, and Rhus, all three of which have been 
doubtfully recorded because of incomplete material. 
In general, the Payette flora indicates a habitat with 

222 R. W. Chaney — Flora of Payette Formation. 

a higher relief than occupied by the Mascall and other 
floras of the Miocene. There is some support for the 
idea of high relief in the coarser phases of the 
Payette formation, though these are decidedly incon- 
spicuous. The bulk of the formation appears to be made 
up of volcanic ash, which may well have been transported 
by wind, along with the leaves of the upland oaks. From 
the nearly uniformly fine character of the Payette sedi- 
ments, a still-water type of deposition may be postulated. 
The fact that there are no evidences of stream deposits, — 
lenticular bodies of gravel and sand, and fossil leaves 
turned into two or three planes during deposition — sup- 
ports the idea of an area of high relief in which the 
streams were eroding rather than depositing. In such 
a region, accumulation of fine sediments and pyroclastics 
would be going on most conspicuously in the basins of 
lakes, into which there would also be blown or washed 
the leaves of trees growing on the slopes above and on 
the borders of the water bodies. In some such way the 
plant-bearing portions of the Payette formation may be 
thought to have accumulated during the Miocene period, 
with a climate much like that of northern California and 
southwestern Oregon today. 

Iowa City, Iowa. 

C. R. Lone/well — Triassie Rocks in Connecticut. '223 

Art. XXII. — Notes on the Structure of the Triassie 
Rocks in Southern Connecticut; by Chester E. Long- 

I. A Group of Small Faults. 

Good exposures transverse to the strike of the rocks 
are Tare in the Connecticut Lowland, and accordingly all 
students of structure in this region have found it neces- 
sary to draw generously on inference in Teaching their 
conclusions. It is therefore especially important that 
every available exposure should be studied and described 
in detail. Dr. H. II. Robinson, Superintendent of the 
Connecticut Geological and Natural History Survey, 
recently called the writer's attention to a tunnel through 
Saltonstall Eidge about four miles east of New Haven. 
This tunnel, constructed in 1901 by the New Haven Water 
Company, was described from the engineer's viewpoint 
by Mr. Edward E. Minor, 1 but apparently its existence 
has not been commonly known among geologists. The 
purpose of the tunnel is to conduct water from Farm 
River into Lake Saltonstall,- an important reservoir, at 
times of protracted drouth. Ordinarily the north end of 
the conduit is kept closed and the floor is covered by only 
a few inches of water. Mr. Minor, General Manager of 
the water company, kindly gave permission for a study of 
the tunnel and furnished equipment for the purpose. Dr. 
Robinson assisted in making the necessary measurements, 
and the writer has discussed the results with him. A 
part of the information for the following description was 
taken from Mr. Minor's paper, referred to above. 

Description of the Tunnel Section. 

The accompanying sketch map (fig. 1) shows the loca- 
tion of the tunnel and indicates the general geology of 
the immediate area. The rocks are Triassie sandstones 
and shales with three interbedded lava sheets. Through- 
out the Connecticut Lowland these rocks have a general 
inclination eastward, but evidently there has been con- 
siderable warping which is responsible for local varia- 
tions in dip and pronounced curvature of surface out- 

1 Proe. Conn. Soe. Civil Eng., 1904, pp. 12-21. 

Am. Jour. Sci. — Fifth Series, Vol. IV. No 21.— September, 1922 

2^4 Longwell — Structure of the Triassic 

crops. Saltonstall Ridge is the surface expression of the 
middle or main trap sheet. Immediately north of Lake 
Saltonstall the ridge curves abruptly eastward, following 
the strike of the rocks, and at the location of the tunnel 
the strike is almost directly east, with the dip toward the 
south. Throughout its length the ridge traces the rim 
of an incomplete structural basin, aptly referred to by 
Davis 2 as a "dish" or " half -boat-like basin," which 
terminates abruptly against the great fault bounding the 
Triassic area on the east. 

The tunnel cuts the ridge in the direction N11°E. 
South of the brick conduit indicated in the section (fig. 2) 
there is no lining except for a few feet in a shatter zone 
between stations 10 and 11. Thus the solid rock is 
exposed practically without interruption for a distance 
of 1800 feet. In the cut near the south portal coarse 
Triassic sandstone is exposed, dipping 35° south. The 
top of the lava sheet is amygdaloidal, and extreme decom- 
position prevents location of the exact contact with the 
overlying sandstone. Inside the portal the amygdaloidal 
rock gives way to massive basalt, and this is succeeded 
by diabase as coarse-grained as that from large intrusive 
sheets such as West Bock. Columnar jointing is evident, 
but not prominent. Near the base of the sheet the dia- 
base again grades into dense basalt, which rests on hard 
conglomeratic sandstone. The surface of contact, 
sharply defined and essentially a plane, is inclined 35° 
to the south. Thus the tilt of the lava sheet may be taken 
as 35°, and its computed thickness, allowing for the 
obliquity of the tunnel, is 475 feet. Possibly this figure 
is slightly in excess of the actual thickness, because of 
small normal faults which are not obvious in the massive 
igneous rock. No evidence was seen that more than one 
now is represented, but such evidence might easily escape 
notice in the narrow section given by the tunnel. 

No metamorphism is apparent below the contact, but 
such effect is not expected, inasmuch as the topmost sedi- 
mentary beds are of conglomerate and coarse sandstone. 
The beds exposed between the lava sheet and the brick 

2 Davis, W. M.: Structure of the Triassic Formation of the Connecticut 
Valley, U. S. Geol. Survey, 7th Ann. Kept., p. 478, 1886; also, The Triassic 
Formation of Connecticut, U. S. Geol. Survey, 18th Ann. Eept., part 2, p. 85, 
1897. '* ' 

Rocks in Southern Connecticut. 


conduit have an average clip of 35° to 36 r southward, and 
the apparent thickness of the section is 535 feet. Prob- 
ably at least 50 or 60 feet should be deducted because of 
duplication by small faults. About one-fourth of the 
section would be classed as fine-drained sandstone and 

Fig 1. 

Fig. 1. — Generalized geologic map of Saltonstall Eidge and vicinity. 
Horizontal lining represents the area of old crystalline rocks. Cross-hatch- 
ing represents the exposed edges of buried basaltic lava flows ; 1, the lower 
or "anterior" sheet; 2, the middle or "main" sheet, forming Saltonstall 
Eidge; 2(a), the same sheet in Totoket Eidge; 3, the upper or "posterior" 
sheet; 4, probably a part of the upper sheet dragged up along the great 
fault. Triassic sediments underlie the areas shown in plain white. 

sandy shale, the remainder as coarse sandstone and con- 
glomerate. Layers with different textures are inter- 
bedded, with no apparent order. In conglomerate layers 
there are numerous well-rounded pebbles of quartz, but 
fragments of granitic rocks, more or less angular, are the 

226 Longwell — Structure of the Triassic 

most conspicuous constituents. The sandstones are dis- 
tinctly arkosic, and have a firm cement of calcium car- 
bonate with some ferric oxide. Throughout the section 
the predominant color is reddish or pink, although some 
of the coarser-grained layers are gray. As exposed in 
the tunnel the beds appear quite regular in thickness, but 
no doubt in more extensive exposures a lenticular char- 
acter would be revealed. The only noticeable variation 
in dip occurs between stations 11 and 12, where the beds 
are distinctly buckled and are almost horizontal for a 
few feet. 

The sedimentary beds are cut by numerous joints, but 
these are not closely spaced except in a few narrow zones. 
Most of the joints strike northeastward. Only those 
fractures on which displacement could be detected were 
studied particularly, and these are indicated in fig. 2. 
They are sixteen in number, including one in the lava 
sheet just above the sandstone. No doubt other small 
faults affect the igneous rock, but in the absence of bedded 
structure it is difficult to find evidence of displacement. 
One of the faults represented in fig. 2 was not studied 
directly, because it is hidden by lining, made necessary 
by the shattered condition of the rock ; but data concern- 
ing it were furnished by Mr. Minor, who was present 
during the construction of the tunnel, and some inferences 
may be drawn from examination of the closely associated 
fractures just outside of the concealed zone. The exact 
strike of this fault is in some doubt. Of the other fifteen, 
ten lie between N60°E and N75°E (averaging N65°E), 
and ^.ve between N35°E and N50°E (averaging N45°E). 
All of the planes (including the one now concealed) dip 
to the northwest, although one is nearly vertical. The 
traces shown in the section are in every case flatter than 
the actual planes, because the tunnel intersects the faults 
obliquely. The true values of the dips, in order from 
north to south, are: 66°, 88°, 58°, 70°, 51°, 68°, 65°, 65° 
(inferred from study of associated fractures, and from 
Mr. Minor's description), 67°, 70°, 65°, 70°, 62°, 76°, 58°, 
72°. The average is 67°. 

The normal character of practically all the faults is 
quite conclusive, and probably this is true of all of them. 
Definite horizons are difficult to locate in the section, 
because beds of similar nature and thickness are repeated 
many times. Observed slickensides are too poorly devel- 

Bocks in Southern Connecticut. 


























:- x 


- O 
> =H 

22S Long iv ell— Structure of the Triassic 

oped to give definite inf ormation, but they at least suggest 
that the lateral component of movement was negligible. 
In most cases the throw is small, ranging from a few 
inches to a few feet. The combined offset effected by the 
ten faults between stations 9 and 11 is determined approx- 
imately, by projecting dips. Before the construction of 
the tunnel began, the surface of the ridge was surveyed 
accurately, and the contact between the lava sheet and 
the sandstone on the north slope was uncovered in order 
to measure the inclination. Therefore the position of the 
contact at the surface as indicated in fi.g. 2 is exact, and 
the offset between this point and the base of the tunnel, 
measured at right angles to the bedding, is 40 feet. Any 
representation of the quantitative distribution of this 
offset between the several faults involved can be sug- 
gestive only, but it is probable that the greater part of 
the throw is effected by one fault, the eighth in order 
from the north. This fault is marked by abundant gouge 
and a shatter zone of considerable width, features that 
are absent or but faintly developed in connection with 
the other dislocations. 

Possible Relationship to Larger Structural Features. 

Of course the section described above cannot be used 
by itself as the basis for generalizations regarding 
regional structure. Minor faults are numerous in the 
Triassic of Connecticut, and many of them doubtless have 
strictly local explanations. Development of joint sys- 
tems may well have attended the warping shown in the 
curvature of Saltonstall Eidge, and minor displacements 
might be expected to occur either contemporaneously with 
the formation of the joints or at a later date. However, 
certain characteristics of the faults in the tunnel are 
highly suggestive. All of the planes strike generally to 
the northeast, all of them dip northwestward, and appar- 
ently all represent normal faulting. It is natural, there- 
fore, to attribute the entire group to a common cause. 
A suggestion that this cause is local warping leads to 
an examination of the Saltonstall "dish" as a unit. The 
map shows a number of faults affecting the basin, but 
all of them appear to strike northeastward. Fractures 
developed as a consequence of the warping would be 

Rocks in Southern Connecticut. ^9 

expected to exhibit some symmetry with respect to the 
curve of the ' ' dish, ' ' and not a common parallelism. The 
faults of the Saltonstall basin have the trend character- 
istic of the region. Davis pointed out that faults in the 
southern part of the Connecticut Lowland extend gener- 
ally from southwest to northeast, 3 and a glance at the 
geologic map of Connecticut confirms this statement. 
Davis and other students of the region have also called 
attention to the predominance of faults with apparent 
downthrow on the west. Thus the small faults at the 
tunnel appear to fit into the regional scheme, and we are 
led to suspect that they are related genetically and chron- 
ologically to the larger faults nearby; in other words, 
that they are "sympathetic faults." 

Our knowledge of the existence of large faults affecting 
the Triassic area depends mainly on the repetition or 
offsetting of trap ridges, and not on actual observation 
of fault surfaces. Because of the character of the evi- 
dence the exact nature of displacements is largely specu- 
lative, and in this connection two especially obvious 
questions are presented. 

(1) To what degree was the displacement lateral, 
parallel to the strike of faults ? 

(2) "What is the direction of dip on the major fault 
planes i 

Some indirect evidence bearing on the first question is 
found for certain parts of the Triassic area. Thus the 
absence of important offsetting in the outcrops of the 
great dikes in the towns of Hamden and Wallingford 
may be taken to indicate that the dikes are essentially 
vertical and that displacement on the large oblique faults 
which evidently extend from the vicinity of Meriden 
southwestward had no important lateral component. 4 
There are grounds for a similar conclusion regarding the 
faults that affect the Saltonstall basin. The map (fig. 1) 
shows that in the northern part of the "dish" the trap 
ridges are offset toward the north on the west side of 
faults, whereas in the southern part the offset is uni- 
versally in the opposite sense. This arrangement — the 
"receding order" and "advancing order" of Percival— 
is exactly that to be expected if all of the faults occurred 

ri 7th Ann. Kept. U. S. Geol. Survey, p. 474. 

4 Davis, W. M. : 18th Ann. Kept. U. S. Geol. Survey, part 2, p. 102. 

230 Longivell — Structure of the Triassic 

with relative downward displacement on the west and 
with no appreciable lateral movement. 

Evidence relating to the second question is very meager. 
There are some facts to indicate that the great boundary 
fault on the east side of the Triassic area dips westward, 
at least in part of its course, and some geologists have 
assumed that the other faults are generally of the normal 
type. A study of the diagrammatic cross-sections in the 
literature, however, reveals striking differences of con- 
ception. The vertical fault-traces in some diagrams evi- 
dently represent efforts to appear noncommittal as to the 
actual attitude ; but a fault which is now vertical may 
well have had an original eastward inclination, if we 
conceive that the blocks suffered some rotation during 
the tilting of the entire Triassic area. In Barren's 
sections 5 most of the faults are shown as reverse, with 
a steep inclination of the planes to the east. Evidently 
this conception is permitted by the surface evidence, 
which merely indicates that in most cases there has been 
relative downthrow on the west. Barrell sought an 
explanation of the general eastward dip of the Connec- 
ticut Triassic in a geanticlinal uplift along the old Taconic 
axis, and conceived that during the uplift subcrustal flow 
from the limb toward the crest might produce shear in 
the overlying crust, breaking the sediments by steep 
reverse faults. As the movement continued the blocks 
would be rotated, and thus some of the faults may have 
been turned to a nearly vertical position. 6 

There is little probability that this question will be 
settled by direct observation. It is possible, however, 
that an accumulation of indirect evidence may lead to 
general acceptance of one view or the other. Groups of 
minor faults which appear to be of the "sympathetic" 
type should give valuable testimony, for they should in 
general imitate the inclination as well as other character- 
istics of the major fractures. It is readily seen that the 
attitude of the small faults at the Saltonstall tunnel favors 
the view of normal faulting for the Saltonstall basin. 
The average inclination determined for these planes (67°) 

5 Barrell, J.: Central Connecticut in the Geologic Past, Conn. Geol. and 
Nat. Hist. Survey, Bull. 23, 1915. 

6 Personal communication. These ideas are also outlined in a manuscript 
left unfinished "by Professor Barrell. 

Bocks in Southern Connecticut. 231 

agrees Well with the average for normal faults in general. 
It is true that if we imagine the beds turned back to a 
horizontal position (or perhaps a few degrees past the 
horizontal, to allow for an original westward dip on a 
bajada slope), most of the faults would then be inclined 
steeply eastward; but this would represent the extreme 
conception that faulting began precisely at the time of 
initial tilting, and even with this allowance the planes 
would be so nearly vertical that they could not be con- 
sidered as the result of ordinary shearing stresses. 

The testimony furnished by this one group of struc- 
tural features is of course only suggestive, and is pre- 
sented here merely as a possible small part of the cumu- 
lative evidence through which we may eventually deter- 
mine, with a reasonable degree of assurance, the 
mechanics of post-Triassic faulting in the Connecticut 

II. Triassic Fax Deposits. 

The prevalence of coarse sediments along the eastern 
border of the Triassic area has attracted the attention 
of all geologists who have worked in the region. Coarse 
conglomerate containing an abundance of fresh feldspar 
is the type of rock found almost universally in outcrops 
immediately west of the great bounding fault. Pebbles 
and bowlders of crystalline rocks included in this con- 
glomerate can be traced to their sources in the meta- 
morphic and igneous formations of the Eastern Highland. 
As a rule the bowlders are fairly well rounded and do not 
exceed a few inches in diameter ; but at certain localities 
adjacent to the fault the fragments in the conglomerate 
are unusually large and many are remarkably angular. 
Some of the fragments belong to exceptional and distinc- 
tive rock types which can be seen in place only in bodies 
of limited size exposed immediately east of the fault. 
The distance to which these coarse deposits were orig- 
inally distributed westward cannot be ascertained, 
because the beds now dip steeply to the east and have 
been bevelled by post-Triassic erosion. Along the strike 
the coarse material at a typical locality has a center of 
maximum thickness and coarseness, grading into finer 
sediments in either direction. This peculiar distribution, 
considered in connection with the rude and lenticular 

232 Longwell — Structure of the Tr lassie 

bedding, the angularity and lack of assortment displayed 
by the fragments, and the evident source of the material, 
leaves no doubt that the coarse sediments were deposited 
in the form of fans at the base of a steep scarp or slope 
by small streams flowing westward. 

A number of these fans are recognized along the fault 
between Branford and Durham. Perhaps the most con- 
spicuous example is exposed in the bluff on the west side 
of Lake Quonnipaug, several miles northeast of the area 
represented in the map, fig. 1. At that locality the angu- 
larity of the fragments making up the fan is truly remark- 
able. Even in the Great Basin the writer has not seen 
products of erosion which show as little evidence of trans- 
portation except in talus deposits, and yet the material 
at Lake Quonnipaug has the rude bedding characteristic 
of coarse fans. We can only conclude that the sediments 
now exposed at that place were deposited directly in 
front of a cliff, and the fragments were supplied by very 
local streams. 

The best example of a fan in the Saltonstall basin is 
located about two miles from Branford in a direction 
slightly east of north. (See fig. 1.) E. 0. Hovey 7 and 
J. D. Dana 8 commented on the unusual character of the 
conglomerate at this locality. Hovey refers to the area 
occupied by the deposits as the "bowlder ridges/' and 
he has described the essential features of the section as 
follows : 

"The region containing the most interesting part of the con- 
glomerate, i. e. the ridges, is about three-fourths of a mile long 
from S. W. to N. E. and perhaps half a mile wide. There are 
several of these ridges each of which is 300 to 400 yards long ; and 
their general trend is N25°E. They are narrow, begin and end 
abruptly, have almost precipitous sides, and rise about 125 feet 
(aneroid measurement) above the brook and meadow bounding 
their region on the east and south. The valleys between them 
are narrow and from 40 to 60 feet deep 

' ' The rock of the ridges is exceedingly coarse. Weil rounded 
bowlders a foot in diameter are very numerous, while others two, 
three, and even four feet long are by no means rare. Much trap 

7 Hovey, E. O. : Observations on some of the trap ridges of the East 
Haven-Branf ord region, with a map : This Journal, 3d ser., vol. XXXVIII, 
pp. 361-383, Nov. 1889. 

8 Dana, J. D. : The Four Bocks, with walks and drives about New Haven, 
p. 191, 1891. 

Rocks in Southern Connecticut. 233 

is present in the rock. In some of the western ridges the con- 
glomerate is mainly composed of trap fragments with a small 
amount of coarse sandstone cement; the middle ridges contain 
many bowlders of quartzite. mica and hornblende schist, gneiss, 
and granite ; while the eastern ridges seem to have rather more 
of the latter constituents than of trap. Bowlders of coarse sand- 
stone also occur in this rock, and many of the fragments of trap 
have coarse sandstone adhering firmly to them and forming a 
part of the bowlders imbedded in the conglomerate. The pieces 
of trap are either angular, subangular, or well rounded, compact 
or amygdaloidal. and quite fresh or much decomposed. Many of 
them contain long, vermiform cavities, either empty or filled with 

calcite But few of the trap bowlders, and those 

widely scattered, occur in the western and southern part of the 
conglomerate area.'' 9 

After careful field study the present writer indorses 
the foregoing description with the exception of one essen- 
tial item — namely, the alleged occurrence of sandstone 
bowlders and of sandstone adhering to trap fragments as 
constituent parts of original bowlders. Obviously the 
presence of such fragments w^ould be of great interest, 
and of considerable importance in interpreting the history 
of the deposits, inasmuch as the only sandstone in the 
region is of Triassic age. Accordingly determination of 
this point w^as made a special object of field study, and 
the conclusion has been reached that the sandstone masses 
mentioned in the description are probably parts of the 
fan matrix, which is composed almost entirely of the 
coarse arkose characteristic of the Triassic sediments in 
the region. Between large angular bowlders the matrix 
commonly has the form of isolated slabs or irregular 
masses, which in weathered outcrops might be mistaken 
for angular pieces originally deposited in the fan. It 
appears that the original constituents w r ere derived only 
from the crystalline formations and fissure veins in the 
Eastern Highland and from masses of basalt. 

The valleys separating adjacent "bowdder ridges " are 
peculiar in that they do not follow the strike of the beds 
but cross it at a large angle. The parallelism of the 
valleys, and the fact that some are not now occupied by 
continuous stream courses, suggest a common and special 
origin. Their trend corresponds closely to the apparent 

9 Hover, E. 0. : loe. cit., p. 375. 

234 Longwell — Structure of the Triassic 

direction of faults recognized to the north and west (see 
fig. 1) and of the great fault immediately southwest of 
the ridges. Accordingly it is suggested that they are 
depressions developed along fault zones. 

The fan deposit is in the upper part of the Triassic 
section as it is known in Connecticut, and it forms the top 
of the section at this particular locality. Its thickness 
certainly is measured in hundreds of feet and may exceed 
a thousand feet near the middle of the fan. The greatest 
degree of coarseness and angularity occurs above the 
upper or "posterior-" lava sheet, although typical fan 
deposits continue to a much lower horizon. It appears 
that the lava flow wedges out in the fan, for there is a 
progressive thinning of the sheet eastward, and no out- 
crops of it can be found in the area of coarsest and thick- 
est fanglomerate. Fragments of basalt, many of them 
vesicular and others quite compact, occur in large 
numbers beneath the "posterior" sheet as well as above 
it. The present dip of the beds is in a direction slightly 
east of south and averages about 35°, although dips as 
high as 45 c are found locally. 

The special interest of the fan lies in its bearing on the 
structural history of the region. In this connection the 
large size and extreme angularity of the fragments are 
very significant features, testifying to the presence of a 
steep scarp near the present border of the Triassic area 
late in the period of sedimentation. Sediments thousands 
of feet in thickness had already accumulated in the Tri- 
assic trough, and without doubt much of this material was 
stripped from the Eastern Highland during long con- 
tinued erosion. The coarse detritus in the fan, therefore, 
furnishes strong evidence of important faulting along or 
near the line of the great boundary fault before the close 
of the Triassic period. 

Abundance of vesicular basalt in the fanglomerate is 
another notable feature. It is to be expected that the 
lava flows would furnish some detritus before their burial, 
but their widespread distribution in the Lowland and the 
apparent absence of important unconformities in the 
Triassic section indicate continued low relief during sedi- 
mentation within the area where the flows now exist. 
Moreover it is evident that the fan materials came from 
the Eastern Highland, and therefore we must conclude 

Rocks in Southern Connecticut. 235 

that lava once covered a portion of the crystalline rocks 
east of the great fanlt. Basalt dikes cutting those rocks 
are not uncommon, and it is quite conceivable that erup- 
tions occurred there contemporaneously with the flows in 
the Lowland. It is also possible that the main sheet 
formerly extended farther eastward. The view is held 
that recurrent faulting maintained steep topography east 
of the area of sedimentation during much of Triassic 
time ; but the scarp may have receded considerably to 
the east during intervals of quiet, and it is not at all 
unlikely that the heaviest flows, aggregating hundreds of 
feet in thickness, encroached somewhat on the crystalline 
area even if the eruptions occurred only in the Lowland. 
With later displacement on the fault the lava on the 
upthrown block would be stripped away, furnishing part 
of the material for the burial of the main body of the 
sheet. Davis's diagrams 10 express the conception that 
the entire Triassic section once extended considerably 
east of the present boundary, with later faulting and 
complete removal of the sediments and lavas from the 
area of the Eastern Highland. There is good suggestive 
evidence, however, that recurrent faulting maintained the 
eastern border of the Triassic area near its present posi- 
tion during a considerable part of the period of sedi- 
mentation. In addition to the testimony of the fanglom- 
erate, discussed here, features and relationships observed 
by Mr. W. L. Russell during recent work along the fault 
zone testify to recurrent faulting and renewal of the scarp 
through the period represented by the section including 
the three trap sheets. 

The essential uniformity of dips in all parts of the 
Saltonstall basin is noteworthy. In the tunnel section, 
discussed above, the dip is 35°, and this figure may be 
taken as the average for the main sheet throughout the 
length of Saltonstall Eidge. The fanglomerate beds, 
which lie more than 1000 feet higher in the stratigraphic 
section, are also inclined toward the boundary fault at 
an angle of about 35 c , although at the time of their deposi- 
tion these coarse deposits must have had a slight inclina- 
tion in the opposite direction. It appears, therefore, that 

10 The Triassic Formation of Connecticut, U. S. Geol. Survey, Ann. Eept 
18, part 2. See especially Plate 20. 

236 C. R. Long iv ell — Triassic Rocks in Connecticut. 

any faulting immediately preceding the fan deposition 
was not accompanied by perceptible tilting of the Triassic 
floor in this area. It is also to be noted that in a complete 
section across the Triassic at this latitude the dips are 
much steeper on the east than on the west. In the vicinity 
of West Rock the eastward inclination is 20° to 25° ; at 
East Rock it is about 20 3 ; between Fair Haven and East 
Haven it ranges from 20° to 30° ; and from East Haven 
to the eastern boundary the average is about 35°. This 
relationship is directly the opposite of that observed for 
the Connecticut Valley as a whole. Davis 11 was the first 
to point out a perceptible decrease in dip from west to 
east, basing his statement on the observed inclination of 
basal beds exposed on both sides of the valley. He inter- 
preted this relationship as favoring the hypothesis of 
downfolding of the Triassic trough during sedimentation. 
Others consider the entire section exposed on the eastern 
side of the area as younger than the basal beds on the 
west, and have cited the decreasing dip from west to east 
as favoring the hypothesis of repeated faulting along 
the eastern side during sedimentation, with progressive 
tilting of the Triassic floor. 12 Another possible inter- 
pretation is suggested here. If the monoclinal tilting of 
the Connecticut Triassic is due to uplift along a geanti- 
clinal axis to the west— and this is a popular conception — 
we should expect a gradual decrease in dip toward the 
east, with increasing distance from the axis of uplift. It 
is seen, therefore, that the relation between dips on the 
two sides of the valley agrees with several different 
hypotheses but actually proves none. However, as dis- 
cussed above, there is other and more direct evidence of 
recurrent faulting on the east. 

A ready explanation of the higher dips on the east in 
southern Connecticut, in contrast with the general rela- 
tionship, is found in the powerful warping reflected by 
the curving of Saltonstall and Totoket ridges. This pro- 
nounced folding, which has materially modified the mon- 
oclinal^ structure, appears only in the southern part of the 
Triassic area. 

11 Davis, W. M., op. eit., pp. 38-39. 

12 Foye, W. G., in a paper on Connecticut structure presented before the 
Geological Society of America, 1921; Bice, W. N., in discussion of Foye's 

E. T. Wherry — Amphisymmetric Crystals. 237 

Art. XXIII.— Amphisymmetric Crystals 1 ; by Edgar T. 


That the symmetry-class to which a given crystal 
belongs is not always evident from external features is 
too well known to require discussion. The methods com- 
monly applied in discovering hidden symmetry may be 
divided into two groups, designated in a general way as 
physical and chemical, respectively. The individual 
methods falling in each of these groups are listed in 
Table I. 

Table I. — Methods of Discovering Hidden Symmetry. 

(A) Physical: 

(a). Excitation of pyroeleetricity. 

(Z>)- Observation of rotation of the plane of polarized 

(c) Study by X-rays. 

(B) Chemical: 

(a) Development of etch-figures by momentary action of 

(&) Production of growth forms by causing rapid crystal- 
lization, by adding impurities to the solution, etc. 

"When the results obtained by applying all of these 
methods agree in placing the crystal in a single symmetry 
class, the relation is merely that of mimicry or pseudo- 
symmetry and requires no further attention. However, 
as pointed out in a previous paper, in a not inconsiderable 
number of cases the different methods do not agree 2 ; and 
for this lack of agreement the term amphisymmetry is 
here suggested. 

In cases of amphisymmetry, crystallographers have in 
general selected as characteristic of the substance one of 
the classes indicated by the special methods of study, and 
have disregarded the other. The most influential factor 
in making the choice appears to have been the circum- 
stance that certain forms are common to crystals of two 
or more symmetry classes, so that if only such forms are 

1 Presented in abstract at the meeting of the Mineralogical Society of 
America, Amherst, Mass., Dec. 29, 1921. 

2 E. T. Wherry, J. Wash. Acad. Sci., 8: 480, 1918; questioned by A. E. 
H. Tutton, ibid., 9: 94, 1919. 

238 E. T. Wherry — Amphisymmetric Crystals. 

developed, a crystal may seem more symmetrical than it 
really is (pseudo-symmetry). Accordingly, when differ- 
ent methods have indicated different classes (amphisym- 
metry) it has been customary to select as typical that 
possessing the lesser degree of symmetry. For example, 
the alkali halides show consistently, when studied by the 
methods of group A, cubic holosymmetry (Class 32) ; yet, 
just because those of group B yield forms with the lower 
cubic gyroidal symmetry (Class 29), this series of com- 
pounds is assigned by practically all crystallographers to 
the latter class. On the other hand, in diamond the 
methods of group A indicate cubic holosymmetry (32), 
those of B hextetrahedral (31). In this case some 
crystallographers follow the same rule as with the halides, 
and assign diamond to the class of lower symmetry, while 
others equally arbitrarily assign it to the higher. 

As long as crystallography was of little interest to any- 
one working in other sciences, the fact that symmetry 
classes were at times selected arbitrarily does not appear 
to have attracted any particular attention ; but after Laue 
thought of passing X-rays through crystals and his 
striking results led physicists to read up on crystallog- 
raphy, considerable perplexity began to be registered. 3 
Thus the Braggs in "X-Rays and Crystal Structure/ f 
p. 157, say: 

"This structure [worked out by X-rays] lias holohedral sym- 
metry. If the copper and oxygen atoms have these exact geo- 
metrical positions, the crystal of cuprite ought to exhibit holohe- 
dral symmetry. Instead of this, certain uncommon forms of the 
costal show that its symmetry is in reality holoaxial [Class 29]. 
If this holoaxial symmetry corresponded to a large distortion . . . 
it is probable that the spectra would show the influence of this 
distortion .... On the other hand the c^stallographic evidence 

3 Since the submission of this paper for publication, Wyckoff (This 
Journal, 3: 177, 1922) has brought out the difficulty in a striking way. He 
urges further study of the crystallography of ammonium chloride (and 
inf erentially of other halides) . It is shown in the .present contribution that 
what is needed is a more discriminating interpretation of existing data ; the 
' ' crystallographic information ' ' supposed to point " to an enantiomorphic 
hemihedry ' ' does not do anything of the sort, but merely shows that the 
atoms when set free from the crystal structure exhibit some enantiomorphic 
features, which their electron arrangement necessitates in any case. Ammo- 
nium chloride, one of the 15 members of the sodium chloride group tabulated 
below, is structurally holosymmetric and only latently gyroidal (enantio- 
morphic) . Compare also footnote 6, below. 

E. T. Wherry — Amphisymmetric Crystals. 239 

would seem to show that the distortion of the structure is slight. 
The etch figures of cuprite indicate a holo symmetric [Class 32] 
crystal, and the crystal does not rotate the plane of polarization 
of light. It is only on account of the existence of crystals which 
show a holoaxial hemihedrism that cuprite is assigned to the 
holoaxial class. 

"These observations apply equally well to the case of potas- 
sium chloride .... assigned by crystallographers to the holoaxial 
class of the cubic system .... examination by means of the X-ray 
spectrometer indicates .... perfect cubic symmetry .... "When 
a potassium chloride crystal is etched with water, it displays 
.... holoaxial symmetry .... Again however there is no trace of 
rotatory polarization. ' ' 

It appears to the writer that in order to avoid misunder- 
standings and reproaches crystallographers should no 
longer make arbitrary choices in cases of amphisym- 
metry, but should face the issue squarely and state both 
symmetries represented in each case. The mode of state- 
ment suggested in his cited paper is, however, capable 
of improvement, and it is the purpose of the present con- 
tribution to further develop the idea. In this connection, 
the writer wishes to express his thanks to Dr. Elliot Q. 
Adams, formerly of this Bureau, for many helpful sug- 

The only reasonable interpretation of amphisymmetry 
which suggests itself is that some of the methods for 
bringing out hidden symmetry yield information as to 
the symmetry of the structure or point-system as it 
stands, while others indicate the symmetry which the con- 
stituent atoms or molecules may exhibit when released 
from the crystal, — in other words, a symmetry latent in 
the atoms. It is therefore in order to consider the signi- 
ficance of the several methods, and the group of alkali 
halides may w T ell be used throughout by way of illus- 

(A) Physical group. — When pyroelectricity is excited 
in a crystal, the opposite charges develop at the ends of 
axes wmich are structurally polar, having different kinds 
of atoms present at their opposite ends. The alkali 
halides yield no pyroelectricity, indicating the absence of 
polar axes in their structure and showing definitely that 
the structures cannot possess the symmetry of the 
gyroidal class. 

Ajl Jour. Sci.— Fifth Series, Vol. TV, No. 21. — September, 1922. 

240 E. T. Wherry — Amphisymmetric Crystals. 

When polarized light traverses a crystal, in any direc- 
tion whatever if it is isotropic, or along an optic axis if 
anisotropic, rotation of the plane will occur if the direc- 
tions of attraction between atoms are to any extent gy- 
roidally arranged, or if, as it is called in stereochemistry, 
" asymmetric " atoms are present. As the alkali halides 
show no trace of rotation of the very sensitive plane of 
polarized light, their structure can certainly not be to the 
least extent gyroidal, nor their atoms asymmetric. 

Since crystals affect X-rays in accordance with the 
spacings between layers of atoms or molecules in various 
crystallographic directions, this method of study permits 
the recognition of the character of the structure. The 
alkali halides behave toward X-rays as if they are in 
every respect cubic-holosymmetric. To be sure, since the 
theoretical rate of decrease in intensity of reflection with 
order is not definitely known, slight deviations might not 
be recognizable, but it is noteworthy that this result 
agrees with those of the more certain methods previously 

(B) Chemical group. — When a solvent is applied 
momentarily to a crystal face, it starts to act at imperfec- 
tions, and develops there minute pits, bounded by more 
or less vicinal faces. Such faces may also be produced 
during the growth of crystals, especially when rapid. 
Vicinal faces may be explained most simply by combining 
the views of the three principal writers on the subject. 4 
Surface tension tends to make crystal surfaces curved, 
while the building up of atoms or molecules in regular 
rows in obedience to the attractive forces acting between 
them tends to fill out all curves or depressions and pro- 
duce plane faces with sharp intersections. The actual 
surface features of a given crystal will be the result of 
equilibrium between these two opposing tendencies, in 
such a way that the vicinal faces actually produced repre- 
sent the maximum deviation from normal angular posi- 
tion which is possible without increasing the growth rate 
sufficiently to cause the filling up of the resulting gaps 
in the surface layers of atoms or molecules. 

In the course of momentary solution as applied in the 
production of etch figures, the symmetry represented in 

4 G. Wulff, Z. Kryst, Min., 34: 449, 1901; H. A. Miers, Phil. Trans. A, 
202: 459, 1903; P. Niggli, Z. anorg. allgem. Chem., 110: 55-80, 1920. 

E. T. Wherry — Amphisymmetric Crystals. 241 

the arrangement of attraction directions around the 
atoms, whether it is exhibited in the structure or is sup- 
pressed by their mode of combination, has a chance to 
make itself felt. In the alkali halides the etch figures 
occasionally show gyroidal symmetry in their bounding 
vicinal faces, indicating this arrangement of the attrac- 
tion directions around one or both kinds of atoms present. 

When there is present, in the solution of a compound, 
an impurity which is attracted to one of the constituents 
in directions other than those perpendicular to the princi- 
pal faces, this impurity will tend to deposit on the grow- 
ing crystal in the crystallographic positions corresponding 
to the attractions and, by preventing the normal constitu- 
ents from building on there, it may produce crystal forms 
other than those typical of the pure substance. It should 
be pointed out in this connection that the relation between 
the attraction directions and the faces developed is the 
opposite of that shown in the stereochemist's diagrams 
of atoms. In the latter, valence lines (attraction direc- 
tions) emerge at apexes of solids, whereas directions 
along which impurities are attracted lie perpendicular to 
the resulting faces. The addition of a large amount of 
magnesium chloride to a solution of one of the alkali 
halides, potassium chloride, is recorded to produce gen- 
eral (hkl) forms of varying indices, probably though not 
certainly gyroidal in arrangement. 5 The magnesium 
may well be attracted to chlorine along force lines per- 
pendicular to (hkl) faces, and prevent some potassium 
from taking up its normal position, so that these faces 
appear on the crystals. This, like the preceding method, 
is thus adapted to bring out the symmetry of the attrac- 
tion directions of atoms. 

The current theory of the constitution of alkali halide 
crystals is that each metal atom has given over an electron 
to a halogen, so that the surfaces of both kinds of atoms 
are completed octets. The structure is then capable of 
becoming relatively highly symmetrical, and the evidence 
in fact indicates cubic holosymmetry. Apparently, how- 
ever, when atoms are dissolving away or when impurities 
are depositing, the single surface electron of the free 
metal, or the seven surface electrons of free halogen, 
being obviously incapable of highly symmetrical arrange- 

5 L. Wulff, Sitzb. Akad. Wiss. Berlin, 1894, 1, 387. 

242 E. T. Wherry — Amphisymmetric Crystals. 

ment, exert attractions in gyroidal fashion. In general, 
it is to be inferred that the methods here grouped as 
chemical may bring out, not the actual symmetry of the 
crystal structure or of the atoms as they stand, but sym- 
metry latent in those atoms. 

The recognition that the various methods of bringing 
out hidden symmetry differ in significance, as above out- 
lined, makes it possible to replace arbitrariness by ration- 
ality in the choice of symmetry class. If it is desired to 
base the classification of a crystal on its actual structure 
or point system, then the physical methods, which bring 
out the symmetry of that structure, should be depended 
upon as indicating the class. If, on the other hand, the 
purpose is to study atomic forces, then the indications of 
the chemical methods, which bring out the latent sym- 
metry of the atoms or molecules, may be followed. X-ray 
workers need consider in general only the structural sym- 
metry, but chemists may be interested more in the latent 
symmetry relations. 

In the alkali halides pyro-electricity, polarized light, 
and X-rays' all agree in indicating the structure to be 
cubic holosymmetric, and they should in general be so 
classified. On the other hand, etch-figures show that the 
constituent atoms have latent gyroidal symmetry, so that 
in any application of the properties of the halides to 
working out the features of alkali metal or halogen atoms, 
they may be assigned to the cubic-gyroidal class. The 
statements concerning amphisymmetric substances in cur- 
rent crystallography books are accordingly inadequate. 
Groth in his "Chemische Krystallographie " states that 
potassium chloride is "Kubisch (pentagonikositetrahe- 
drisch)," i.e., cubic (gyroidal). A better statement 
would be : " Cubic ; structurally holosymmetric ; latently 
gyroidal" (or, if class numbers are preferred, "structur- 
ally of class 32, latently 29"). To cite an entirely differ- 
ent example, Dana in his "System of Mineralogy" 
describes scheelite, calcium tungstate, as "Tetragonal; 
with pyramidal hemihedrism. ' ' In this case neither 
pyroelectricity nor polarization-plane rotation could 
occur, so that this statement represents all that could be 
ascertained about the mineral at the time. However, the 
"hemihedrism" (or merosymmetry) is shown only by 
rare faces, probably produced by films of impurities, and 

E. T. Wherry — Amphisymmetric Crystals. 243 

recent X-ray study has shown the structure to be holo- 
symmetric, making it now possible to be more specific: 
"Tetragonal; structurally holosymmetric ; latently 
pyramidal. ' ' 

In conclusion, there is presented (Table II) a list of 
substances which give evidence of being amphisymmetric, 
with their respective structural and latent symmetry 
classes, as far as these are at present known. Where a 
group of compounds shows similar relations throughout, 
only one typical member is given, but the number of 
members included is indicated by numerals in parentheses 
after the mineral name. The order is in general that of 
increasing complexity of the compounds and of decreas- 
ing symmetry. Class names and numbers are modified 
from Groth's "Physikalische Krystallographie. " 

Table II. Amphisymmetric Substances. 

Mineral name. 
Sodalite gr.(4) 
Halite gr.( 15) 
Xitrobarite gr. (3) 

Cobaltite gr.(5) 
Scheelite gr.(4) 
Scapolite gr.(5) 

Aragonite gr.(4) 




Na 4 Al 3 aSi 3 12 

KA1(S0 4 ),.12H,0 

Ba(N0 3 ), 
NaSrAs0 4 .9H,0 
K 2 Mg 2 (S0 4 ) 3 
NaAlSi0 4 
Cu g A1C1SO 10 .9H,O 
CaW0 4 

Na 4 Ca 4 Al 9 ClSi 15 49= 
TiTi0 4 
Pb,Cl,C0 3 
NiS0 4 .6H 2 
KH,P0 4 gr. 
PbMo0 4 
Ag 5 SbS 4 
CaC0 3 

AgN0 3 

H,Na,P,0 6 .6H,0 
LLS0 4 .H,0 
Sr(C,H,6,)(N0 3 ). 
liH 2 

Symmetry Classes. 
Structural. Latent. 

Cubic holosym. (32) 
Cubic holosym. 
Cubic holosym. 
Cubic holosym. 
Cubic holosym. 
Cubic holosym. 
Cubic holosym. 
Cubic hextetrah.(31) 
Cubic hextetrah. 
Cubic hextetrah. 
Cubic diploidal( 30) 
Hexag. holo.(29) 
Hexag. holo. 
Tetragonal holo. (15) 
Tetragonal holo. 
Tetragonal holo. 
Tetragonal holo. 
Tetragonal holo. 
Tetragonal holo. 
Tetragonal holo. 
Orthorh. holo. 
Orthorh. holo. 
Orthorh. holo. 
Orthorh. holo. 
Monoclinic holo. (5) 
Monoclinic holo. 
Monoclinic holo. 


C. hextetrah. (31) 
C. hextetrah. 
C. hextetrah. 
C. diploidal(30) 
C. diploidal 
C. gyroidal(29) 6 
C. gyroidal(29) 8 
C. tetartoidal(28) 
C. tetartoidal 
C. tetartoidal 
C. tetartoidal 
Hex. pyr.(23) 
Ditrig. pyr.(20) 
Tetr. bipyr.(13) 
Tetr. trapez. 
Tetr. trapez.(12) 
Tetr. trapez. 
Tetr. trapez. 
Tetr. scaleno(ll) 
Orth. pyr. 
Orth. pyr. 
Orth. bisphen.(6) 
Orth. bisphen. 
Monocl. domat. (4) 
Monocl. domat. 
Monocl. sphen.(3) 

Asymmetric (1) 

6 It is noteworthy that since all of the supposedly cubic-gyroidal substances 
are amphisymmetric, no crystal is known which structurally belongs to this 

244 E. T. Wherry — Amphisymmetric Crystals. 


The fact that different methods may not agree in their 
indications as to the symmetry class of a given crystal 
is re-emphasized, and the term amphisymmetric is pro- 
posed for this relation. Physical methods show the sym- 
metry of the structure, chemical methods bring out latent 
symmetries of the separate atoms or molecules. It is 
recommended that in stating the crystallization of an 
amphisymmetric substance the two classes indicated be 
given, one designated structural, the other latent. 
Finally, a table of a large number of substances which 
give evidence of being amphisymmetric is presented. 

Bureau of Chemistry, 

U. S. Department of Agriculture, 
January, 1922. 

T. H. Clark — New Trilobite Appendage. 245 

Art. XXIV. — A New Trilobite Appendage; by Thomas 

H. Clark. 

Among all of the faunal assemblages in the Levis 
formation the most interesting and the richest in hitherto 
unrecognized fossils is the so-called Shumardia limestone. 
Billings referred to this as ''thin-bedded limestone inter- 
stratified in the graptolitic slates" (Pal. Foss. Can. 1865, 
p. 93) in recording the locality of the single new fossil 
which he described from it. That only one species should 
have been known at that time is very remarkable, for in 
places it is crowded with fossils of all descriptions. To 
be sure, the limestone is difficult of access in place, and it 
may be that Billings was not content with fallen blocks 
at the base of the cliff. T. C. Weston, lithologist for the* 
Canadian Geological Survey, later made large collections 
from this horizon, which he named the Shumardia lime- 
stone on the labels which he affixed to his specimens. Dr. 
Raymond recognized the name in his paper on the succes- 
sion of the faunas at Levis, and listed under his Zone D, 
the fauna then known to characterize that bed. Eecently 
the writer has made a large collection of this horizon 
which includes all of the previously recognized species, 
and a great many new ones. An incomplete summary 
of the fauna, which is not yet thoroughly worked over, is 
as follows : — 

Plants Obscure algse. 

G-raptolites 19 species. 


Inarticulate 11 species. 

Articulate 2 species. 


Trilobites 10 species. 

Ostracods Numerous. 

Cirripedea 2 species. 

Annelids (?) 

Conodonts 4 species. 

Serpulites 1 species. 

With the exception of a few small articulate brachiopod 
shells, all of the specimens of the species listed above are 
now composed of black carbonaceous material, which indi- 
cates that their original shell or skeleton was of chitin or 

216 T. H. Clark — New Trilobite Appendage. 

of some similar organic compound. This holds true for 
the fossils of the black shales with which the Shumardia 
limestone is interstratified. Without doubt the fauna of 
the shale and the limestone was a pelagic one, well 
adapted to a floating or a swimming existence, and pre- 
served in such abundance in the rocks of the Levis forma- 
tion because of the absence of any indigenous benthonic 
predaceous or scavenging forms. The abundance of car- 
bonaceous matter in the shales, and the absence of any 
truly benthonic forms, indicate that the bottom of the 
Levis sea was not capable of sustaining a fauna; this 
would account for the preservation, almost in their 
entirety, of many very delicate graptolites and thin- 
shelled brachiopods. 

One of the most remarkable fossils from this horizon 
is a single detached appendage of a trilobite. It is small, 
less than 5 mm. in length and about 1 mm. wide. It 
consists of the outer part of a coxopodite, a large exopo- 
dite, a fragmentary endopodite, and one or two other 
members to be discussed later. The accompanying figure 
shows the disposition and relations of these parts. It 
should be borne in mind, however, that this figure repre- 
sents the writer's interpretation of the specimen. The 
proximal parts are somewhat crushed, so that other inter- 
pretations are possible. 

Fig. 1. — Sketch of the trilobite appendage from the Shumardia limestone, 
Levis formation, Levis, Quebec. X 12. 

a. Exopodite. b. Endopodite. c. Coxopodite. d. Second coxopodite. 
e. Second exopodite. f. ?Ventral membrane. 

The exopodite is by far the largest and the most prom- 
inent member. Its outline is as shown in the sketch, blunt 
at the inner end, terminating in a long spine at the distal 
extremity, widest beyond the middle. The posterior 
margin carries a score or so of setae, which are short, 
stout and pointed outwards. These appear to be inserted, 
in the margin of the exopodite, rather than outgrowths 

T. H. Clark— New Trilobite Appendage. 247 

of it, for the border is thickened and the setae appear to 
come ont from nnder it. Moreover, adjacent setae, none 
of which appears to be flexible, emerge from the border 
at varying angles, a condition which could scarcely obtain 
were they prolongations of the substance of the exopodite. 
Inserted setae on trilobite appendages have hitherto been 
found upon endopodites only. The thickened border 
extends along the posterior side of the exopodite, which 
has been crushed fiat, as is shown by the many cracks 
and slight displacements. 

The endopodite is incomplete. Two or three segments 
are shown, but it does not appear that the whole member, 
if completely restored, would be as much as 3 mm. long ; 
it probably would not exceed 2 mm. It is circular in 
section, and was undoubtedly stout, for the thickened 
border of the exopodite may be seen to bulge above it. 

The coxopodite, adjacent to the blunt end of the exopo- 
dite, is small, but probably incomplete. It is roughly 
rectangular, and crushed flat. 

Below this are two fragments which the writer inter- 
prets as a second coxopodite, and the base of a second 
exopodite, both belonging to a single segment. Above 
the whole is a large patch of black, but less shiny, mate- 
rial, shown by shading in the sketch, which, if it be part 
of this assemblage at all, might represent a fragment of 
the ventral membrane of the trilobite. This is purely 
conjectural, but it lies where the ventral membrane would 
be looked for, immediately above the exopodite. 

Apart from the presence of inserted setae in the exopo- 
dite, the most remarkable feature about this specimen is 
the relation in size between the exopodite and the endopo- 
dite. That the animal which bore this appendage was 
not adapted for crawling is certain ; no trilobite with such 
reduced endopodites could have been a successful ben- 
thonic inhabitant. On the other hand, the large paddle- 
like exopodite might well have been adapted to a nectonic 
or a planktonic mode of existence. This is in accord with 
the conclusion reached with regard to the whole Levis 

This limb does not agree exactly with any of the known 
trilobite appendages. In Neolenus the endopodites are 
relatively very large, much longer than the exopodites. 

248 T. H. Clark — New Trilobite Appendage. 

In the Asaphidae, no exopodites are definitely known, but 
the endopodites are long*, although slender. In Triar- 
thrus, the basipodites, ischiopodites and meropodites are 
for the most part much modified in form, but in any case 
the endopodites are as long as, if not longer than, the 
exopodites. In the form of the exopodite alone, some 
comparison might be made between the present specimen 
and Walcott's figured specimen of Ptychoparia cordil- 
lerae Rominger. 1 In all other trilobites the exopodites 
are no longer than the endopodites. Thus, off anatomical 
grounds alone, it would be impossible to connect this form 
with any trilobite whose appendages are known. Of the 
genera represented in the Shumardia zone only Isotelus 
and Triarthrus have so far yielded recognizable appen- 
dages. The most abundant trilobite is Shumardia granu- 
losa Billings, which outnumbers all others about five to 
one ; but this species is much too small to have possessed 
the appendage in question. The remaining genera repre- 
sented in the Shumardia limestone are Agnostus, Endym- 
ionia, Symphysurus, Telephus, and Holometopus, to any 
one of which this appendage might belong. There seems 
to be no evidence for placing it within any particular one 
of these genera. 

Museum of Comparative Zoology, 
Cambridge, Mass. 

1 Smithson. Misc. Coll., vol. 67, pi. 21, figs. 3-5, 1918. 

J. V. Lewis and L. H. Bauer — Cyprine. 249 

Art. XXV. — Cyprine and Associated Minerals from the 
Zinc Mine at Franklin, N. J.; by J. Volney Lewis and 
Lawsox H. Bauer. 

The following notes refer chiefly to cyprine, the sky- 
bine variety of vesuvianite, in intimate microscopic mix- 
ture with willemite, and to the associated rhodonite (var. 
bustamite) and andradite (var. polyadelphite). 

1. Vesuvianite, var. Cyprine. — Blnish green fibrous 
cyprine was found in granite in the Parker shaft, at 
Franklin, in 1905, and was described by Professor 
Palache, 1 who also published the analysis by Steiger, 
which is quoted below for comparison. 

The mineral here described was found in the ore body 
near the hanging wall in a crosscut 374.5 feet south and 
361.5 feet west of the Parker shaft, and 10 feet above the 
850-foot level. It is sky-blue in color ; in texture it varies 
from fine granular and fibrous to dense. To the naked 
eye, and even under the hand lens, it appears to be homo- 
geneous. The analysis (A, below) showed nearly 22 per 
cent, zinc oxide, and since the oxide ratios fell within the 
limits of variation of vesuvianite, it was supposed that 
a new variety of this mineral had been found. 

Thin sections showed, however, that while vesuvianite 
is the dominant mineral, it is plentifully sprinkled with 
rounded grains and hexagonal crystals of willemite, with 
dimensions up to .16 mm. in diameter. Measurements 
by the Rosiwal method showed that this mineral consti- 
tutes 29 per cent, of the volume, or 33.3 per cent, of the 
mass. If the zinc, iron, and manganese of the accom- 
panying analysis (A) are assigned to willemite, the cal- 
culation gives 32.68 per cent, of this mineral. The 
remaining constituents, calculated to 100 per cent., repre- 
sent approximately the composition of the vesuvianite. 
A few minute particles of metallic copper were visible 
under the microscope, but doubtless most of the copper 
determined is in combination. 

The analysis of this mixture and the results of the 
calculations are given here, together with Steiger 's 
analysis 2 of cyprine from the Parker shaft. 

1 Charles Palache : Contributions to the Mineralogy of Franklin Furnace, 
this Journal, (4), 29, p. 184, 1910. 

2 U. S. Geol. Survey, Bulletin 591, p. 315, 1915. This statement of 
Steiger 's analysis, which is followed here, differs slightly from that given 
by Professor Palache. 

250 J. V. Lewis and L. H. Bauer — Cyprine. 

Analyses of vesuvianite from Franklin, New Jersey. 

(A) . (B) (C) (D) 

Si0 2 32.42 8.70 35.14 36.41 

A1 2 3 14.07 20.86 17.35 

FeO 77 .77 1.86 

MnO 1.50 1.50 1.75 

ZnO 21.71 21.71 1.74 

CuO 99 1.47 1.48 

CaO 25.22 37.40 33.21 

MgO 1.08 1.60 1.38 

H 9 2.38 3.53 (— .24 


Others .... 1.15 a 

100.14 32.68 100.00 100.08 

A. Vesuvianite-willemite mixture, apparently homogeneous. 
L. H. Bauer, analyst. 

B. "Willemite calculated from A. 

C. Remainder of A, recalculated to 100 per cent. — approxi- 
mately the composition of the vesuvianite. 

D. Steiger's analysis of vesuvianite from the Parker shaft. 
Sp. gr. 3.451, carefully freed from specks of metallic copper. 

2. Rhodonite, var. Bustamite. — Pale pink in color; 
elongated cleavable to coarsely fibrous in texture. Cal- 
cium replaces manganese to a remarkable degree, as shown 
by the analysis (E), by L. H. Bauer, with which is com- 
pared (F) the mineral from Langban, Sweden, as given 
by Dana (System of Mineralogy, p. 380) : 

Analyses of bustamite from Franklin, N. J. and Langban, 


(E) (F) 

SiO, 46.72 47.66 

Al 2 6, 1.34 

FeO 46 .48 

MnO 26.51 31.65 

ZnO 1.34 

CaO 22.24 18.16 

MgO 1.27 1.18 

Others .98 b 

99.88 100.11 

aPbO tr.; Na 2 .44; K 2 .50; F .36; less = F .15. 
bBaO .19; alkalis .27; gangue .52. 

J. V. Lewis and L. H. Bauer — Cyprine. 251 

3. Andradite, var. Polyadelphite. — The brown gran- 
ular garnet associated with the vesuvianite approaches 
typical andradite in composition (analysis G, by L. H. 
Bauer) more nearly than the specimens from Franklin 
represented by the older analyses, (H) and (I), quoted 
by Dana : 

Analyses of Andradite from Franklin, N. J. 

(G) (H) (I) 

SiO, 34.28 34.83 33.72 

Al.,6. 3.12 1.12 7.97 

Fe.,O a 25.53 28.73 17.64 

MnO 7.41 8.82 16.70 

CaO 29.20 24.05 25.88 

MgO 39 1.42 (Ign. .08) 

99.93 98.97 101.99 

Besides the minerals named above, brown and reddish 
brown phlogopite is also abundant in scales and crystals, 
and in places coarse cleavable feldspars. Some of the 
latter give extinction angles corresponding to labradorite 
and others to anorthite. There are also very rare grains 
of pyrite, a small amount of cleavable calcite, and a little 
dark green pyroxene. 

October 26, 1921. 


I. Miscellaneous Scientific Intelligence. 

1. Stratigraphy of Northwest Greenland; by Lauge Koch. 
Med. dansk geol. Forening, 5, 1920, No. 17, 78 pp. map, 1 : 2,500,- 
000 ; sections, views and diagrams ; bibliography of 29 titles. — 
The Second Danish Thule Expedition to North Greenland in 
1916-18, headed by K. Rasmussen, included Lauge Koch as geolo- 
gist, who has presented a valuable summary of his results. A 
great part of the region consists of a plain of Archean gneiss, 
sloping gently to the northwest, where it is covered by nearly 
horizontal early Paleozoic strata, 400 meters in maximum thick- 
ness, consisting of conglomerates, sandstone, slate and limestone, 
apparently of shallow-water formation. Non-disintegrated 
feldspar grains in the coarser strata and a prevailing red color 

252 Scientific Intelligence. 

suggest that they represent deposits eroded from a neighboring 
part of the Archean plain under arid conditions. The fossils 
include brachiopods, graptolites and corals. In strong contrast 
to this undisturbed region, the northwest coast is occupied by a 
well folded and greatly degraded and dissected mountain chain, 
presumably of Devonian deformation, which is continued 
obliquely across Robeson and Kennedy canals into Grinnell land. 
Koch regards this mountain belt as a curved extension of the 
Caledonian folding of northwestern Ireland, Scotland, western 
Norway, and western Spitsbergen. w. m. d. 

2. Revue de Geologie et des Sciences connexes. — This Review, 
of which the third year began with Januaiy, 1922, is a monthly 
publication issued under the patronage of the Societe Geologique 
de Belgique at Liege, Belgium. It deserves the support of all 
geologists and mineralogists. Subscription for 1922 fifty francs ; 
address the General Secretary, Laboratoire de Geologie, Univer- 
site de Liege. 

3. First Pan-Pacific Commercial Conference. — This import- 
ant conference has been called by the Pan-Pacific Union to meet 
at Honolulu on October 25. A wide range of topics has been 
announced for discussion, the sessions extending until October 31. 
Following this date a week will be given to a series of attractive 
excursions, the return to San Francisco being scheduled for 
November 8. 

4. First Congress of Industrial Chemistry. — The first Chem- 
ical Exposition, organized by the Societe de Chimie Industrielle, 
will be opened at the National Conservatoire des Arts et Metiers 
on October 7. The program already announced is a guarantee of 
an important and valuable meeting. 


Dr. Alexander Graham Bell, the inventor of the telephone, 
died on August 2 in his seventy-sixth year, at his estate at Beinn 
Bragh near Baddeck, Nova Scotia. Born in Scotland, he came to 
this country at the age of twenty-two, and in 1876 his first tele- 
phone patent was granted. What the telephone to-day, in its pres- 
ent form, means to the activities of the civilized world is too well 
known to need remark. The story of his active, useful life, with 
its many signal achievements besides the one to which his name 
has been so long attached, is picturesque and remarkable and will 
always be of general interest. The many honors he received were 
not more than his contributions to science and humanitv merited. 

Ward's Natural Science Establishment 

A Supply-House for Scientific Material. 

Founded 1862. Incorporated 1890. 

A few of oiir recent circulars in the various 
departments : 

Geology: J-32. Descriptive Catalogue of a Petrograpbic Col- 
lection of American Eocks. J- 188 and supplement. 
Price-List of Rocks. 

Mineralogy: J-220. Collections. J-238. Minerals by Weigbt. 
J-224. Autumnal Announcements. 

Paleontology: J-201. Evolution of tbe Horse. J-199. Palae- 
ozoic index fossils. J-115. Collections of Fossils. 

Entomology: J-dd. Supplies. J-229. Life Histories. J-230. 
Live PupaB. 

Zoology : J-223. Material for dissection. J-207. Dissections 
of Typical Animals, etc. J-38. Models. 

Microscope Slides: J-189. Slides of Parasites. J-29. Cata- 
logue of Slides. 

Taxidermy: J-22. Nortb American Birdskins. Z-31. General 

Human Anatomy: J-37. Skeletons & Models. 

General : J-228. List of Circulars & Catalogues. 

Ward's Natural Science Establishment 

84-102 College Ave., Rochester, N. Y., U. S. A. 


number consisting of 100 to 120 pages). Editor : EUQENIO RIQNANO. 

This is the only review which has a really international collaboration ; which is 
of world-wide circulation and occupies itself with the synthesis and unification of 
knowledge, in the history of the sciences, mathematics, astronomy, geology, physics, 
chemistry, biology, psychology and sociology. 

It is the only review which, by inquiries among eminent scientists and writers 
(on: The philosophical principles of the sciences; Fundamental astronomical and physical 
questions of current interest; Contributions given by the various countries to different 
branches of knowledge; Question of vitalism; Social question; International questions 
raised by the world war), makes a study of the problems interesting scientific and intel- 
lectual circles throughout the world. 

It has published articles by Messrs. : Abbot -Arrhenius -Ashley - Bayliss- Beichman - 
Bigourdan - Bohlin - Bohn = Bonnesen - Borel - Bouty - Bragg - Bruni = Burdick - Carver - 
Caullery - Chamberlin > Charlier = Claparede - Clark - Costantin - Cromtnelln - Crowter = 
Darwin - Delage - De Vrles - Durkheim - Eddington - Edgeworth - Emery = Enriques - 
Fabry - Flndlay - Fisher = Fowler = Qolgi = Gregory - Harper - Hartog - Heiberg - Hlnks - 
Hopkins=Inigues-Innes=Janet-Kaptein=Kaye=Kidd-Langcvin=Lebedew -Lloyd Morgan = 
Lodge - Loisy = Lorentz =■ Loria -Lowell - MacBride = Meillet -Moret-Muir-Peano -Picard - 
Poincare - Pniseux = Rabaud = Rey Pastor - Righi = Rignano= Russell-Rutherford -Sagnac = 
Sarton- Schiaparelli - Scott - See - Sherrington = Soddy - Starling- Svedberg- Thomson • 
Thorndike-Turner-Volterra -Webb -Weiss -Zeeman and more than a hundred others. 

"SC1ENTIA" publishes its articles in the language of its authors, and joins to the 
principal text a supplement containing the French translations of all the articles that 
are not in French. (Write for a Specimen Number to the General Secretary of 
ti Scientia ,, f Milan, sending, -to defray postal and other expenses. -2 Francs in 
stamps of your country.) Annual subscription : 40 sh., or 10 dollars post free. 

Office : 43 Foro Bonaparte, Milan, Italy 

Publishers: WILLIAMS & NORG ATE - London ; FELIX ALCAN- Paris 



Art. XVII. — The Determination of the Space Group of a 

Cubic-Crystal; by R. W. G. Wyckoff, 175 

Art. XVIII. — The Symmetry and Crystal Structure of Zinc 
Bromate Hexahydrate, Zn(Br0 3 ) 2 .GH 2 0; by R. W. G. 
Wyckoff, 188 

Art. XIX. — On the Symmetry and Crystal Structure of 
Sodium Hydrogen Acetate, NaH(C 2 II 3 2 ) 2 ; by R. W. G. 
Wyckoff, 193 

Art. XX.— Cone-in-Cone; by W. A. Tarr, , 199 

Art. XXI. — Notes on the Flora of the Payette Formation; 

by R. W. Chaney, 214 

Art. XXII. — Notes on the Structure of the Triassic Rocks 

in Southern Connecticut; by C. R. Longwell, 223 

Art. XXIII. — Amphisymmetric Crystals; by E. T. Wherry, 237 

Art. XXIV.— A New Trilobite Appendage; by T. H. Clark, 245 

Art. XXV. — Cyprine and Associated Minerals from the Zinc 

Mine at Franklin, N. J. ; by J. V. Lewis and L. H. Bauer, 249 


Miscellaneous Scientific Intelligence. — Stratigraphy of Northwest Greenland, L. 
Koch, 251.— Revue deGeologie et des Sciences connexes: First Pan-Pacific 
Commercial Conference: First Congress of Industrial Chemistry, 252. 

Obituary.— A. G. Bell, 252. 

Library, U. S. Nat. Museum. 


OCTOBER, 1922 

Established by BENJAMIN SILLIMAN in 1818. 




Editor: EDWARD S. DANA. 



op Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 

of Washington. 



No. 22-OCTOBER, 1922. 


19 22. 


Published monthly. Six dollars per year, in advance. $6.40 to countries in the 
Postal Union ; $6.25 to Canada. Single numbers 50 cents; No. 271, one dollar. 

Entered as second-class matter at the Post Office at New Haven, Conn., under the Act 
of March 3, 1879. 


Now Ready 

Elements of Optical Mineralogy. 


Second Edition, Entirely Eewritten and Much Enlarged. By ALEXANDER 
N. WlNCHELL, Doct. Univ. Paris, Professor of Mineralogy and Petrology, 
University of Wisconsin. 

The treatment of Principles and Methods given in this new edition is 
adapted to the microscopic study of minerals, both ill thin sections 
and in powdered form immersed in liquids, the latter methods 
permitting an accurate quantitative measurement of the index of refraction. 

216 pages. 6 by 9. 251 figures. Cloth, $3.50. 

Other Wiley Books for Mineralogists 

Determinative Mineralogy: With Tables for the Deter- 
mination of Minerals by Means of Their Chemical and 
Physical Characters. Third Eevised and Enlarged Edition. 


298 pages. 5£ by 8. 31 double-page tables, 81 figures. 



Rocks and Rock Minerals. 

By the late LOUIS V. PlRSSON. 

414 pages. 5 by 7£. 36 half-tone plates. Cloth, $3.50. 

Eemember, any "WILEY book can be obtained on Free Examination 
terms. Send for the ones in which you are interested. 


432 Fourth Avenue 
New York 

London Montreal, Quebec 

CHAPMAN & HALL, Ltd. Renouf Publishing Company 

AJS 10.22 




Art. XXVI. — Jones's Criticism of Chamberlin's Ground- 
work for the Study of Megadiastrophism; by T. C. 

In the June number of this Journal there appeared an 
article by William F. Jones entitled "A Critical Review 
of Chamberlin's Groundwork for the Study of Megadias- 
trophism," 1 which calls for analysis and comment. 

Doubt at once arises as to the real subject of the arti- 
cle, for immediately after the title there follows, without 
an intervening word, the heading : i ' Summary of Thesis, ' ' 
under which five subjects not of the nature of critical 
reviews are outlined, while at the close of the "Introduc- 
tion" immediately following, the purpose of the paper is 
given in these words : 

"The purpose of the present paper is to bring together in 
summary form from the various sources the evidence leading to 
the conclusions that the earth passed through a molten stage and 
that there exists at no great depth a zone which yields to diastro- 
phic differential stresses." 

Immediately following the first paragraph of the intro- 
duction concerned with formal statements, the discussion 
opens thus : 

"It is evident that the study of megadiastrophism is to be one 
of the application of wedge-dynamics to continental areas. To 
consider continental areas as the upper surfaces of deeply point- 

1 Diastrophism and the Formative Processes. XIV. Groundwork for the 
Study of Megadiastrophism. Part I. Summary Statement of the Ground- 
work already laid: Thomas C. Chamberlin. Part II. The Intimations of 
Shell Deformation : Eollin T. Chamberlin. 

This paper is essentially a digest of thirteen previous papers and of sev- 
eral collateral articles. Part I consists of 77 propositions or specifications 
giving- as briefly as practicable the essential features of the fuller papers. 
Part II embraces certain intimations as to the downward extension of 
deformative action drawn from studies on accessible features by a method 
which deduces the under-configuration and deeper phases of the deformed 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 22. — October, 1922. 

254 T. C. Chamberlin — Jones's Criticism of 

ing wedges implies, of course, that the earth lacks an astheno- 
sphere or yield zone" (p. 394). 

As if the critical review of a digest of thirteen long arti- 
cles, the treatment of the five themes of a thesis, and the 
bringing together in summary form from the various 
sources of the evidences that support a molten stage and 
an asthenosphere, afford insufficient scope, the author 
adds a prophecy of what i ' the study of megadiastrophism 
is to be " and his interpretation of what that means. The 
sweep, however, is made wider still; the fundamental 
issues of the schools are set in array, as follows : 

"This postulate is fundamental. The issue is thus sharply 
drawn between the two schools of thought. The basic problems 
of diastrophism, isostasy, and vulcanism must be attacked and 
interpreted in two diametrically opposite ways by the two 
schools. The one school, as represented by T. C. and K. T. 
Chamberlin, interpret these problems from the viewpoint of an 
earth which has maintained perfect solidity through the forma- 
tive eras and down to the present The other school, as 

represented by Daly and the late Barrell, interpret these prob- 
lems from the viewpoint of an earth which passed through a 
stage of fusion" (pp. 394-5). 

To deal with this frontal screen may be a little tedious 
but it is necessary to clear the air if any accurate practice 
is to follow. And first comes the prophecy. Part II of 
the paper on the "Groundwork" touched upon wedge- 
dynamics and it is a fair inference that wedge-dynamics 
will play some part in the superstructure built upon the 
groundwork and an equally fair inference that it will play 
the same sort of part that it does in the inquiry cited in 
the paper. But instead of shaping his forecast critically 
on these lines, the reviewer introduces a characterization 
of a radically different sort. The right-minded reader 
will want to know : 

What are the intimations of ic edge-dynamics as set 
forth in Part II of the paper on " 'Groundwork." The 
purpose of Part II was to call attention to the fact that 
there is available a working method of using the intima- 
tions of the accessible deformations in determining the 
downward extension of deformative action, and to bring 
this extension into service as a part of the start already 
made in the study of the larger questions of diastrophism. 

Groundwork for Study of Megadiastrophism. 255 

The prime function of the method is to disclose the actual 
zones of weakness, or at least zones of yield, that is, one 
type of asthenozones. Though the matter dealt with was 
solid, it was not "perfect solidity" in the sense implied 
by the reviewer's statement. Obviously, as the diagrams 
and text of the original papers abundantly show, it was a 
yielding solidity. In its application to the Appalachians 
of Pennsylvania and the Colorado Rockies, the method 
disclosed plunging zones of yield, but at the same time the 
masses enclosed within these yield tracts showed the same 
sort of deformation that had usually been assigned them 
by advocates of a molten earth. The wedge-body did not 
show ' k absolute rigidity" or imply the ''non-existence 
of a yield-state" "for almost any condition of stress" 
(p. 396). The plunging zones of yield are supplemental 
asthenic zones, and add just so much of detectable yield 
to what was known from the accessible deformations 
above. Thus, far from being inhibitory of horizontal 
yield — the assigned function of the asthenosphere — the 
plunging yield-zones do good service by revealing the 
mode by which horizontal yield actually took place in 
these particular cases. 

The reed ciuestion of an asthenosphere in the earth as it 
actually is. — Our critic's definition of the asthenosphere 
as a zone at no great depth "which yields to diastrophic 
differential stresses" is so broad as to include any view 
that accepts general internal deformation. Under the 
new view of the constitution of matter, it seems clear that 
all zones yield to diastrophic differential stresses. The 
real question is whether there is a zone at no great depth 
below the surface which yields in a special degree com- 
pared with the zones above and below. 

The doctrine of an asthenosphere as set forth by its 
author, Dr. Joseph Barrell, was made to relate mainly to 
a solid elastico-rigid zone, as will be seen by reference to 
the original papers. 2 He assigned a mode of movement 

- J. Barrell, ■ ' The Strength of the Earth 's Crust. ' ' Part VI. Kelations of 
Isostatie Movements to a Sphere of Weakness — the Astlrenosphere : Part 
VII. Variations of Strength with Depth as Shown by the Xatnre of Depar- 
- from Isostasy. See. A. Presentation of Theory. Sec. B. Applica- 
tions of the Theory: Part VIII. Physical Conditions Controlling the 
Nature of Lithosphere ami Asthenosphere. Sec. A. Belations between 
Eighlity, Strength, and Igneous Activity. Sec. B. Belations with other 
Fields of Geophvsics. Jour. Geol., vol. 22, pp. 655-'-:;. 729-41. 1914. vol. 
23, pp. 27-44, 425-32, 499-515, 1915. 

256 T. C. Chamberlin — Jones's Criticism of 

of the same nature as that of glaciers, as interpreted by 
Chamberlin (VIII, p. 433), and that of " solid rock-now, " 
as interpreted by Van Hise (VIII, p. 428). In dealing 
with this question, then, on the lines laid down by its 
author, we have nothing directly to do with the molten 
state of the earth. If any indirect effect is inherited from 
the molten state, it is merely the favorable or unfavorable 
condition of the solid matter it has left. The practical 
question of to-day is merely the alternative whether 
a solid zone formed by congelation from a molten state 
would be more, or would be less, favorable to yield than 
a zone formed from minute mixed solid accretion. To me 
it seems that the accretional mixture offers somewhat the 
greater susceptibility to yield in glacier or solid-rock- 
flow fashion, but if someone else would give different 
values to the respective susceptibilities, it need not put us 
into "diametrically opposite" camps. So also in the 
matter of continental warping, Barrell placed the zone 
at 50 to 500 miles below the surface. It seems to me that 
such a small direct up-lift — of the wedging type if you 
please — as the relief of the continents implies, would be 
as little likely to destroy the iveakness of the zone as an 
equal warp by lateral crumpling or any other mode of 
deformation. Neither would seem to me quite fatal to 
the weakness of the zone. 

The assignment of 'personal attitudes. — The reviewer's 
sweeping introductory statement gives the impression 
that R. T. Chamberlin and T. C. Chamberlin are sharp 
antagonists of the doctrine of an asthenosphere and are 
about to promulgate a scheme of wedge-dynamics that, if 
successful, will "of course" be fatal to it. As a matter 
of fact, E. T. Chamberlin has been accustomed to teach his 
classes that an asthenosphere, in the sense of a special 
yield-zone, is at least a possibility. As his special method 
of investigating mountain diastrophism is designed to 
show where the yield-zones actually are — so far as 
implied by the features of shell deformation — and as his 
studies are no more than fairly begun, he naturally 
remains merely a hospitable student of the doctrine of an 
asthenosphere and refrains from propagandism until the 
evidences of the case return their verdict. I am not now 
teaching and I have refrained from going into print on 
this subject. Instead, I have been trying to gather the 

Groundwork for Study of MegadiastropJvism. 257 

import of the new concept of matter and to make use of 
that as a guide in the study of this and other phases of 
megadias trophism, as indicated in my last serial article. 3 
Under the new dynamics of matter, it seems clear that the 
greater diastrophisms must affect the whole globe in due 
proportions, but these proportions are far from being 
equal at all depths. The question of special degrees of 
diastrophism in particular zones remains. I have been 
trying to balance the considerations that favor special 
diastrophism in the zone of magma tic generation against 
the considerations that favor special diastrophism in the 
zones of least resistance near the surface, but I have not 
closed the study and have not furnished any matter on 
this particular topic for the reviewer to review. 

The reviewer's intimation of sharp antagonism seems 
to do Dr. Barrell even more injustice than us. In a letter 
(Dec. 3, 1913) transmitting to me, as Editor of The Jour- 
nal of Geology, the first part of his manuscript on "The 
Strength of the Earth's Crust,"— in the Vlth, Vllth, and 
Vlllth articles of which he makes his initial statement of 
the doctrine of an anesthenosphere — he says : 

"I am sending you the first four parts before the fifth is 
completed because I just read this past week the series of papers 
on Diastrophism and the Formative Processes which you are 
beginning to publish. My paper falls so much in line with that 
subject that I think it might be considered as a part of the 

Nor, in suspending my series and giving precedence to 
Barrell's — because the two could not well run simul- 
taneously, in fairness to other contributors and to readers 
more interested in other subjects — did I feel that there 
was anything antithetical in the doctrines we were advo- 
cating, though they might differ in sub-features. I think 
the reader will not find in the three original articles on 
the asthenosphere any intimation that a planetesimal con- 
stitution of the zone of weakness was in any respect what- 
ever inhibitory of an asthenosphere. Our critic does the 
doctrine of asthenosphere poor service when he ties it up 
with the theory of a molten stage. Nor is the service 

3 Diastrophism and the Formative Processes: XV. The Self -Compression 
of the Earth as a Problem of Energy, Jour. Geol., 29, pp. 679,700, Nov.- 
Dec, 1921. 

258 T. C. Chamberlin — Jones's Criticism of 

improved by listing reserved and cautious students of the 
doctrine as its fundamental and necessary opponents. 

Where then does the sharp issue really lief — During the 
last century and before, a molten earth was almost uni- 
versally accepted by geologists. It was a logical infer- 
ence from a self-developing or monoecious planetary 
evolution, whatever its particular form. During the first 
decade of the present century, an origin by means of a 
dioecious, or cooperative, or bi-parental evolution was 
proposed, because the conviction had been reached that no 
monoecious system of evolution could give rise to the 
singular dynamic features presented by our planetary 
system. This postulated cooperative evolution was such 
as to give rise to a growth of the earth essentially in a 
solid state by means of the slow ingathering of planetesi- 
mals. There thus arose a radical difference between two 
classes of postulates because of their cosmological inheri- 
tances. The practical difference between them centered 
in a molten vs. a solid state of the earth. Along these 
lines, there is a fundamental issue. 

As a sub-phase in the development of the accretional 
view, an attempt has been made to cut off the top of the 
standard planetesimal tree, so to speak, and graft on the 
stump a molten state instead. I have not specifically dis- 
cussed, in print, the congeniality of such engrafting, and 
I am reluctant to do so now — for personal reasons too 
obvious to need mention — but it is made necessary. 

7s a molten state compatible with a planetesimal origin® 
— The planetesimal basis is herein taken for granted. If 
the conditions of formation thus postulated permitted 
free speculation, it might be assumed that the planetesi- 
mals were gathered in fast enough to force a molten state. 
But the postulated conditions are far from permitting 
free speculation ; they are rather severely restrictive as 
we shall at once see. 

A molten state under planetesimal conditions could 
normally arise only when the mass of the nucleus was 
large enough to give great heat by its own condensation 
and, at the same time, had strong enough attraction to 
draw in the planetesimals rapidly, and even then the 
planetesimal orbits must be favorably distributed. There 
is no gain by supposing that the planetesimals grew into 
large masses, for (1) there would be some loss of energy 

Groundwork for Study of Megadiastrophism. 259 

and much loss of time in such aggregation, (2) what is 
gained in the mass would be lost in the longer intervals 
between infalls, and (3) the energy of infall largely takes 
the form of mechanical dispersion. The larger meteor- 
ites melt less proportionally than the small ones. The 
great infall that produced Meteor Crater (Coon Butte) 
gave a most impressive illustration of the largeness of the 
energy dispersed in the form of side thrusts, of lateral 
scattering of material, of outward radial dispersion by 
the violent reactive explosion that followed the impact, and 
of cooling in the'act of dispersion as shown by the vesicu- 
lar state. On the other hand, the meagerness and evanes- 
cent nature of the melting was equally surprising. 4 

According to the planetesimal hypothesis, the material 
for the formation of the earth was erupted from the sun 
under the stimulus of a passing body which drew the 
projected matter into orbital courses while in its flight. 
To be conservative let us assume that the erupted mass 
of solar gas had only twice the diameter of the present 
dense earth. The central mass of this matter would nor- 
mally be projected vertically or radially from the sun's 
center, and at the same time, as it issued from the sun, 
would be subject to gaseous expansion, while it would also 
be affected by eruptive dispersion and by internal motions 
which would have a dispersing effect in general. The val- 
ues of the expansional and dispersive effects are matters 
of estimate, but it is easy to follow rigorously the vertical 
projection of the central part. The distance projected 
from the surface of the sun was roundly 215 times the dis- 
tance of the erupted mass from the center of the sun when 
it started ; hence the diameter of the erupted mass when it 
reached its orbital distance would have been 430 times the 
diameter of the present earth. To catch the meaning of 
this, we must now consider the power of such a mass to 
control itself against the competing attraction of the sun, 
for this power of self-control determined the size and 
mass of the collecting nucleus. No power of self-control 
would develop in the erupted mass until after it had 
passed outside the Roche limit of the sun. At the dis- 
tance of the earth the sphere of control of the present full- 

4 Diastrophism and the Formative Processes : XIII. The Bearing of the 
Size and Rate of Infall of the Planetesimals on the Molten or Solid State 
of the Earth, Jour. Geol., vol. 28, pp. 686-95, Nov.-Dec, 1920. 

260 T. C. Cliamberlin — Jones's Criticism of 

grown earth is 194 diameters of the earth, The sphere 
of control of the self-controlled portion of the erupted 
mass obviously could not have been so large as this. 
Taking the ultra-maximal sphere of control as a working 
basis to give a margin of safety, the volume of the erupted 
mass due to radial dispersion simply — 430 times the 
earth's diameter — compared with the volume of the pres- 
ent sphere of control — 194 times the earth's diameter — 
shows that less than one-tenth of the erupted mass lay 
within even the ultra-maximal sphere of control. To this 
dispersion should be added the dispersive effects of gase- 
ous expansion and of eruptive scattering. Over against 
these additional dispersive factors there is only the allow- 
ance to be made for the inward curvature of the projected 
constituents due to their mutual gravity during their 
swift short flight under the projectile forces. 

I think that every competent reader will see that, with- 
out laying stress on any factor that depends upon personal 
judgment, the collecting nucleus would only have included 
a minor fraction of the projected matter. The rest must 
have gone into independent orbits from which collection 
by orbital dynamics would have been very slow and the 
development of a holpmolten state of the earth quite out 
of the question. 

It thus appears clear to the point of practical demon- 
stration that a molten state would not arise normally 
under the terms of the planetesimal hypothesis, and that 
the attempt to graft such a state on the hypothesis merely 
gives rise to an incompatibility, not to an issue. 

Contrasted views of rigidity. — It has thus taken much 
space to clarify the issues projected in the Introduction 
before any semblance of real review began, but another 
such projection immediately follows relative to the mean- 
ing of rigidity, as follows : 

"The term 'rigidity' lias been generally loosely interpreted. 
Rigidity is dependent not only on the stress applied but upon 
the time the stress is applied. A material may be rigid for one 

set of stress conditions and non-rigid for another Rigidity, 

then, is merely a relative term. ' ' 

There is just enough of truth and of misconception in 
this to make it very misleading. Two things are easily 
confounded: (1) the property of rigidity whose values 

Groundwork for Study of Megadiastrophism. 261 

are usually given by geo-pliysicists in terms of "constants 
of rigidity,'' and (2) the effects of stresses brought to 
bear on bodies possessed of this property, which may be 
as various as the imagination chooses to make them. In 
considering the effects of stresses of course the time of 
continuance of the stress is important. But before enter- 
ing upon discriminations, let us clear away the impres- 
sion the reader is quite likely to get from the stress laid on 
the alleged variability, that here is something of import- 
ance not known to us — and perhaps not to the reader : In 
the very article under criticism and on the very subject of 
the relations of the more mobile, more deformed parts 
of matter undergoing diastrophism to the less mobile, 
less deformed parts, under conditions of differing depths, 
E. T. Chamberlin says : 

"No limiting depths can be assigned, for the time-element 
plays an important part, thongh not easy to evaluate. To quick- 
acting stresses the earth reacts as an elastico-rigid body ; under 
long-continued stress it yields to slow movement. With greater 
depth molecular rearrangement and recrystallization should pre- 
sumably take precedence" (p. 418). 

Xow the vital question is not what might happen under 
imagined conditions of stress, but what is the real state of 
the earth's matter as revealed by three independent lines 
of evidence interpreted in the light of the stress-conditions , 
time-conditions, pressure-conditions, temperature-condi- 
tions, and other conditions that actually exist, in great 
variety, in the earth. 

This state of the earth substance is what I have empha- 
sized in my papers. I have referred to the state of the 
internal matter as elastico-rigid — or some equivalent of 
this double phrase — with a uniformity and persistency 
that has made it wearisome to my friends and myself 
because such reiteration seemed necessary to avoid the 
ambiguities that cling* about the colloquial term "rigid- 
ity. ' ' I have tried to use the compound term strictly as 
a definition of the properties imparted to the matter 
under consideration by its constitution. I think this is at 
least not at variance with established scientific usage. At 
any rate, my use is definite and requires recognition in a 
critical review, whether approved or not, The important 
point is to keep clear the distinction between the elastico- 

262 T. C. ChamberUn — Jones's Criticism of 

rigid state and the viscous state. The essential consid- 
erations in the former state are fixed attachments of the 
constituents to one another at particular points by a force 
which resists displacement, attended by a reacting elastic 
force which tends to restore any displacement that may 
take place. There is thus implied a resisting-yielding- 
restorative combination. In the viscous state displace- 
ment also requires force but displacement is not followed 
by restorative action. The elastico-rigid state is not, as 
its name clearly indicates, a state of ' i absolute ' ' rigidity, 
but one of elastic yield. It thus becomes the basis for 
certain typical modes of deformation and even of continu- 
ous movement under differential stress, as in the case of 
glacial motion and solid rock-flow as now interpreted by 
the most critical students of these fields. 

In my serial paper on the self -compression of the earth 5 
— which, as my only paper as yet built on the "ground- 
work," should be taken as the chief index as to what "the 
study of megadiastrophism is to be" — I have cited the 
lucid illustration of this property offered by the electro- 
magnetic poles developed by the revolution of electrons 
about positive nuclei. Attachment by minute intense 
poles of this type affords an ideal picture of fixation at 
particular points conjoined with a restorative force. 
This is probably the actual mode of formation of the 
elastico-rigid state, but as it may be premature to insist 
upon this, let it serve merely as a clear-cut index of the 
type of view held. 

Now this is a very different picture from that which my 
critic paints for me. As already stated he seems to me to 
confound the effects of possible stresses not necessarily 
connected in any way with the organization of the matter, 
with the properties of the matter arising from the mode 
of its organization. A spring balance is an illustration 
of an elastico-rigid mechanism, but the properties of the 
balance belong to one category while the varying strains 
of commodities put on it to be weighed belong to a differ- 
ent category. A modern steel bridge is a rigido-elastic 
structure, but examining engineers are accustomed to 
report its properties in fixed terms, not in variables 
dependent on the stresses of a hand-car or a freight train 

5 Diastrophism and the Formative Processes : XV. The Self -Compression 
of the Earth as a Problem of Energy, Jour. Geol., vol. 29, pp. 679-700, 1921. 

Groundwork for Study of Megadiastrophism. 263 

that may happen to be crossing it. Expert geophysicists 
give the earth-qualities under discussion in terms of 
"constants of rigidity" and "constants of elasticity." 

The summary characterization of ChamberVnvs views. 
— Immediately following his exposition of rigidity from 
"the relative viewpoint," and without quoting a sentence 
from the paper announced as being under review, or 
adducing evidence of any other kind, the reviewer pro- 
ceeds to characterize mv interpretations as follows (p. 

"Now Chamberlin's interpretation of the evidences of this 
rigidity — that is rigidity shown under tidal stress and seismic 
vibrations, implies that this state of rigidity and elasticity holds 
good for stress differences of all magnitudes and of all time 
durations. Under this interpretation the terms rigidity and 
elasticity become practically absolute instead of dependent. But 
such an interpretation is essential if he is to postulate the non- 
existence of a yield state anywhere within the earth for almost 
any condition of stress" (p. 396). 

This is a very remarkable statement. It does not repre- 
sent, or even resemble, any thing stated in my paper on 
the groundwork for the study of diastrophism, or in any 
of my diastrophic papers. I have not the slightest 
thought of postulating "the non-existence of a yield zone 
anywhere within the earth for almost any condition of 
stress." On the contrary I postulate yield everywhere. 
My views of yield and movement in elastico-rigid matter, 
in the mooted case of glacial motion, have been outstand- 
ing for more than twenty years. 6 My views on such 
action in megadiastrophism are definitely indicated and 
concretely illustrated in my serial paper next following 
that on the ' ' Groundwork. ' ' 7 By reference to that paper, 
it will be seen (1) that the need of a revision of funda- 
mental ideas of the constitution and behavior of matter in 
accordance with the new disclosures is put to the front; 
(2) that these imply a dynamic organization as open rela- 
tively as a planetary system and hence responsive to all 
stresses ; (3) that these revelations afford a concrete con- 

6 "Glacial Studies in Greenland," Presidential Add., Geol. Soc. Am., vol. 
16, pp. 209-14, 1895. 

7 The Self-Compression of the Earth as a Problem of Energy, Article XV 
of the series Diastrophism and the Formative Processes, Jour. Geol., vol. 29, 
pp. 679-700, Xov.-Dec, 1921. 

264: T. C. Chamberlin — Jones's Criticism of 

cept of the yielding nature of the elastico-rigid state of. 
matter; (4) that the relations of crystalline organization 
to atomic and molecular organization is more intimate 
than appeared nnder the old view and hence the crystal- 
line state is more fundamental and presumably more per- 
vasive than heretofore supposed; (5) that the proportion 
of heat energy in the earth to 4;ne organizing and main- 
taining energies has been greatly exaggerated; (6) that 
the new concept of an open dynamic structure clears away 
many of the seeming obstacles to idiomolecular action; 
and (7) finally, a concrete sketch is given of the way in 
which idiomolecular readjustments act under relatively 
low stresses in producing deformative movement without 
involving the liquid state. The whole working concept 
has the nature of indefinite capabilities of compression, 
deformation and metamorphism by means of a system of 
interchanges facilitated by the. open revolutional mechan- 
ism of the constituent atoms, and molecules. It is there- 
fore responsive to every force brought to bear upon it. It 
is a yield system par excellence and yet a resisting system 
of the dynamic type with fixed forms. With such a basal 
view, it is impossible for me to entertain such notions of 
intractability as those assigned me. 

Substitution of part for the whole. — It seems necessary 
to take notice of another phase of this statement relative 
to my views, to wit : 

"Now Chamberlin 's interpretation of this rigidity — that is 
the rigidity shown under tidal stress and seismic vibrations, 
implies," etc. 

This naturally carries the impression that my interpre- 
tation of rigidity has been based solely upon tidal stresses 
and seismic vibrations, whereas in the very first specifica- 
tion of the paper professedly under critical review, I men- 
tion in addition to tidal and seismic evidence, l ' nutation 
and collateral evidences." The following from the paper 
just cited shows the way in which I usually put my view : 8 

"Tidal 9 and nutational 10 evidences concur in indicating a 

8 Diastrophism and the Formative Processes : XV. The Self -Compression 
of the Earth as a Problem of Energy, Jour. Geol., vol. 29, pp. 691-92, 1921. 

9 A. A. Michelson and Henry G. Gale: "The Eigidity of the Earth," 
Jour. Geol., vol. 27, pp. 585-601, 1919. 

10 W. Schweydar : ' ' Ueber die Elastizitat der Erde, ' ' Naturwissenschaf - 
ten, Part 38. Potsdam, Germany, 1917. 

Groundwork for Study of Megadiastrophism. 265 

higher degree of rigidity and elasticity in the interior, taken as 
a whole, than in the outer shell. Seismic waves add very specific 
confirmatory evidence, so far as the outer seven-eighths of the 
volume of the earth is concerned. The seismic evidence for the 
remaining central part is as yet obscure, and is differently inter- 
preted by the special students of the subject. In a general way, 
the whole of the interior is covered by the tidal and nutational 
evidences. These favor the interpretation of the central part 
as highly rigid and elastic, since these qualities fit the general 
import of the evidence, but for the present it is prudent to leave 
the question of the state of the center to be settled in the future. 
It is to be observed that the increasing density of the interior 
tends to dampen the speed of the seismic waves, and that cor- 
rection for this effect must be made in deducing the hrward 
increase of rigidity and elasticity from the seismic records^ 
"When allowance is made for this, the generalization that rigidity 
and elasticity are notably higher in the interior than in the outer 
shell is put beyond serious question. ' ' 

Xow a special significance attaches to nntation in the 
combination of evidences because much emphasis has 
been laid on the shortness of the seismic and tidal vibra- 
tions and the possibility that they might not mean much. 
The Chandlerian cycle of the nutation has a period of 
about 14 months (432.8 days according to Sclrvveydar). 11 

Now I think that the reader will agree that, in view of 
this putting of a part for the whole and of the assignment 
of ideas of w r edges and rigidity quite foreign to my own, 
I am entitled to disclaim any responsibility for such other 
views as my critic may assign me in the rest of the paper, 
and that it is best that I should use what space remains 
available in discussing some of the more vital issues 
raised. Foremost among these is the question, rendered 
acute by the criticism: 

What is really implied by the properties of rigidity and 
elasticity disclosed by seismic waves, the body tides, the 
earth's nutation and associated evidences^ 

The elastic rigidity disclosed by the seismic waves. — 
Setting aside the earth's nucleus for later consideration, 
there is agreement among seismologists that the outer 
portion of the earth to a depth of 3000 kilometers or more 
transmits transverse or distortional seismic waves and 
that this implies an elastico-rigid state. So far as the 

11 Op. cit. 

266 T. C. Chamberlin — Jones's Criticism of 

seismic vibrations themselves are concerned, they simply 
show the presence of the elastico-rigid property under 
existing conditions. They tell nothing directly as to 
what might or might not be the state of the interior under 
other conditions. As a first step toward reaching the 
wider meaning we must consider what the existence of 
the disclosed state of rigidity and elasticity signifies 
under all the heat, pressure, differential stress, and coop- 
erative influences that now affect the outer seven-eighths 
of the earth's body. We must bear in mind further that 
these conditions are inherited from a previous chain of 
conditions that ran back as far as geologic inheritances 
go, and that this brings in the effects of the time factor. 
Now looking at the matter from the positive or construc- 
tive point of view, the case may be put this way : 

The outer seven-eighths of the earth is to-day elastico- 
rigid — to a degree not determined by the seismic waves — 
notwithstanding all the heat, all the pressure, and all the 
differential stresses within it, together with all cooperat- 
ing effects, as also any help that may have been inherited 
from past geologic time. 

Or, from the negative or destructional point of view, 
the case may be put in this way : 

All the differential stresses in the outer seven-eighths 
of the earth's body, all the heat and pressure in this part, 
and all cooperating effects, together with any help that 
may have been inherited from past geologic time, have 
not proved sufficient to destroy the elastic and rigid prop- 
erties of this part of the earth. 

Now these existing conditions are not simply those that 
prevail at some one horizon but those that exist in all 
parts of the outer body of the earth. They undoubtedly 
include stress-differences of quite different orders of 
magnitude, some of which are presumably high. Their 
significance, therefore, is not to be escaped by academic 
assumptions about dependence. 

The seismic waves have a peculiar value in that they 
search out the special states of matter in the various 
parts of the interior. Their cumulative evidences have 
now practically shut out the molten state from serious 
consideration in present-day problems, such as the 

Groundwork for Study of Megadiastrophism. 267 

asthenozones, the loci and methods of magmatie gene- 
ration, etc. 

The rigidity and elasticity disclosed by the body 
tides. — The body tides are simply the yielding and res- 
toration of the earth in response to differential stresses 
from the moon and snn. The quick restoration to form 
implies a high state of elastic rigidity. The evidence in 
this case differs from that of the seismic waves in that 
the tidal oscillation shows that the earth's action as a 
whole is that of a highly elastic rigid body under existing 
conditions. The periods of action are also longer than 
those of the earthquake vibrations but still relatively 

Now, as before, the body tides show the existence of the 
properties of rigidity and elasticity in the earth acting 
as a whole in spite of all effects of the earth's present 
temperatures, pressures, differential stresses and inher- 
ited effects. 

The rigidity and elasticity disclosed by the nutation of 
the earth. — Nutation may be regarded as a third line of 
testing or sampling the rigidity and elasticity of the earth. 
Like the tides, it discloses the response of the earth as a 
whole to differential stresses, but it supplements the testi- 
mony of both tides and seismic vibrations in an important 
way because its changes of stress are slow and the oscil- 
lations long, the period of the Chandlerian cycle being 
about 14 months ; besides which there is an annual cycle, 
and some deviations that imply other slowly changing 
stresses. Now the significance of the properties of 
rigidity and elasticity under these slower stress actions 
is to be interpreted, as before, in the light of all existing 
temperatures, pressures, and stress differences notv 
affecting the earth. 

AVe have thus three lines of independent testimony, 
differing in kind and in period, and yet none of them has 
revealed the limits of the elastic and rigid properties. It 
falls to anyone who assumes any specific limits to give the 
evidence of such limits, or frankly acknowledge that his 
assumptions are simply speculative. 

By correlating these three lines of evidence with other 
data, as has been done by Schweydar, rather definite 
quantitative values may be reached. Schweydar con- 
cludes that the constant of rigidity of the earth is 3 X 10 11 

268 T.C. Chamberlin — Jones's Criticism of 

dynes at the surface and 30 X 10 11 dynes at the center, 
with a mean rigidity of 2% times that of steel. 12 

What follows the rigid-elastic state when the strain 
limit is passed? — It has now been amply shown by pro- 
tracted researches on glacier motion, solid rock-now and 
dynamic metamorphism, that while the viscous state and 
even liquefaction may, in particular cases, follow when 
the strain limit is reached, and also that faulting, granu- 
lation, or massive shear may follow in others, the more 
widely prevalent, more characteristic and more funda- 
mental sequence, when the pressure is not over-intense 
and the time is ample, is the passage of one elastico -rigid 
state into another elastico -rigid state. The new state is 
often more perfectly and highly organized in respect to 
the elastico-rigid property than the previous state. Thus 
movement may go on to great lengths, in time, simply by 
this succession of solid states, as in the case of glacial 
flow and dynamic metamorphism. The earth seems to be 
progressively passing into more and more highly organ- 
ized states of elastic rigidity, or in other words, meta- 
morphism is in progress in the solid earth. This change 
seems to be effected largely by idiomolecular action, i.e. 
individual action atom by atom or molecule by molecule. 
The main mass remains solid while the shifting atoms or 
molecules act as individual elastico-rigid organizations. 
And further, even this individual action is not necessarily, 
and perhaps not normally in the depths, liquid action, nor 
gaseous action, for liquids and gasses are assemblages of 
molecules, while these shifting atoms or molecules act 
individually. As such individuals, they are, under the 
new views, elastico-rigid mechanisms of an ideal type. 
In being individually detached from its hold on one crys- 
tal or particle, the atom or molecule is probably shot 
through the dynamic — not bodily — resistances of the 
space lattices or equivalent opennesses, until it is arrested 
and oriented by a new fixed attachment. The rigido- 
elastic arrangements of the countless atoms and mole- 
cules in the solid earth body are held to be giving way as 
differential stresses rise to the requisite degree, and thus 
to be passing from old to new relationships. Thus a 
"flow" like that known to affect glaciers and rocks under- 

12 W. Schweydar ; ' ' Tiber die Elastizitat der Erde, ' ' Naturwissenschaf ten, 
Part 38, p. 22, 1917, Potsdam, Germany. 

Groundwork for Study of Megadiastrophism. 269 

going dynamic metamorphism is probably in constant 
progress in the earth. This, however, does not seem to 
be adequate to relieve all increases of stress, and so 
accumulations arise and bring on special stages of dias- 
trophism of a more vigorous type. 

The state of the earth nucleus. — So far, to avoid com- 
plication, the state of the earth's nucleus has been neg- 
lected. As already stated in a quotation from my last 
paper, 13 it is necessary to correct the velocities of the 
seismic waves for density to give the full state of rigidity 
and elasticity. If such correction is applied to the best 
data now available as to the velocities of the seismic 
waves, it shows a rise of elastic-rigidity as far toward the 
center as good records go. The necessity for this correc- 
tion seems to have been overlooked by those who have 
inferred the existence of a state of non-rigidity near the 
center. If question is raised about the trustworthiness 
of the correction, it may be recalled that the mean density 
is much higher than the surface density and that the pre- 
cession of the equinoxes implies higher density in the 
deep interior. To satisfy these approximately, correc- 
tion may be made by use of the classic law of density of 
Laplace, or still better, it may be made by means of the 
formula of Roche which was devised especially to give a 
distribution of internal density in as close accord with the 
astronomical requirements as possible. 

The dying out .of the distortional waves when the 
chords get so long as to cut the radius about half way 
from the surface to the center has been much stressed 
recently as though it were evidence that rigidity ceases at 
about that depth, but the ways in which both transverse 
and longitudinal waves become feeble and die out as the 
chords lengthen, puts another aspect on the matter. 
They must die out sooner or later in the nature of the 
case. It is remarkable that they are able to traverse as 
long chords as they do. The fact that both kinds of 
waves die out in traversing chords much short of the 
diameter of the earth seems to point to their natural 
death by progressive exhaustion. The brachistochronic 
chord of 140° is about the extreme limit of penetration by 
the longitudinal waves, while the similar chord of 110°- 

13 Diastrophism and the Formative Processes : XV. The Self -Compression 
of the Earth as a Problem of Energy, Jour. Geol., vol. 29, pp. 679-700, 1921. 

Am. Jour. Sci.— Fifth Series, Vol. IV. No. 22.— October, 1922. 

270 T. C. Chamberlin — Jones's Criticism of 

120° is about the limit for the transverse waves. Both 
kinds give increasingly imperfect records as their pene- 
trations become great, but not equally so. The longi- 
tudinal waves have a speed about fifty per cent greater 
than that of the transverse waves, evidence that they are 
actuated by a more effective elasticity — that of volume — 
than are the distortional waves, which depend on elas- 
ticity of form. It seems logical, therefore, that the longi- 
tudinal waves should maintain themselves longer than the 
transverse waves. As a matter of fact they do not retain 
their integrity as much longer than the transverse waves 
as might be expected from their superior velocity. They 
grow feeble and doubtful progressively as do the distor- 
tional waves, and at the depth where the latter cease to 
be identifiable they have already become dubious. If the 
distortional waves were strong and gave good records 
uniformly up to a certain depth and then suddenly ceased, 
while the longitudinal waves continued on and gave good 
records up to 180°, the inference of a change from the 
elastico-rigid to a viscous state would at least be plausi- 
ble, but as the case now stands, such an interpretation 
does not seem to be the natural one. This does not 
invalidate the view of Oldham that there is .a change in 
the nature of the material at about the half -radius depth, 
for this is compatible with a solid state. 

The correction of the seismic records for density brings 
the seismic evidence into general harmony with the tidal 
and nutational evidences and makes the presumption 
strong that rigidity and elasticity increase from the 
surface to the center as held by Schweydar. 14 The fol- 
lowing quotation from Walker, a seismologist who had 
unusual experience with instruments and records, lies in 
the same line : 

"At the present time the evidence in favor of a solid earth is 
very great but the alternative view that the interior of the earth 
is fluid retarded for a considerable time the progress of seismo- 
logical theory which requires the earth to possess the properties 
of an elastic solid. As astronomical theory agrees with the seis- 
mological in demanding a solid earth, we accept this as a primary 
condition." 15 

14 W. Schweydar: "tiher die Elastizitat der Erde," Naturwissenschaf- 
ten, Part 38, 1917, Potsdam, Germany. 

13 G. W. Walker, Modern Seismology, p. 13. (Longmans, Green and Co., 
London, New York, Bombay and Calcutta, 1913.) 

Groundwork for Study of Megadiastrophism. 271 

The separation and transmission of seismic vibra- 
tions. — My critic makes the claim that the texture of an 
accretional earth would not be capable of transmitting 
seismic vibrations, and that an " isotropic" medium is 
necessary for the separation of the vibrations. Specific- 
ally he says : 

' ' The two types of waves travel at different velocities but can 
only become distinctly separated out in a homogeneous medium; 
that is. homogeneous as to stress effects, or, in other words, 
isotropic. The resultant vibrations which travel circumferen- 
tially from the shock center pass through what we know is a 
heterogeneous medium. The wave types are not separated in 
the earth's surficial shell' 7 (p. 398). 

The waves here spoken of as "resultant vibrations," 
and elsewhere .as "undifferentiated," are, as I under- 
stand, the long waves. The medium necessary to the 
separation is later spoken of as "a non-crystalline iso- 
tropic material similar to 'undercooled' liquids" (p. 399). 

I understand the facts and their natural interpretations 
to be these : 

That the seat of the earthquake, the point of origin of 
all the seismic waves, lies within the shell not very far 
below the surface; that the vibrations generated there 
are very heterogeneous and mixed in the utmost confu- 
sion; that these vibrations are propagated at different 
speeds strictly because of their natures and their rela- 
tions to the elasticity of the transmitting earth substance ; 
that they are separated as a necessary effect of their dif- 
ferent velocities; that the separation is not dependent on 
any mysterious virtue of an isotropic medium or any 
'undercooled' liquid; that they separate in any medium 
that is capable of transmitting them; that there is no 
special separation but merely a progressive separation as 
long as they keep going; that all known types are trans- 
mitted through the shell and undergo progressive separa- 
tion while being so transmitted ; that nothing is known of 
waves that cannot be transmitted through the shell ; that 
the longest, most abundant, most conspicuous, best 
defined, most notably organized, most far-penetrating 
and scientifically most interesting waves, the undae longas 
and the coda, are transmitted by the shell; that these are 
sometimes so well transmitted that they are recorded at 

272 T. C. Chamberlin — Jones's Criticism of 

stations from both circumferential directions, one of 
which of course is usually more than 180° from the point 
of origin; that the primary (longitudinal) and secondary 
(transverse) waves are very inconspicuous features of 
the seismograms as such — though extremely interesting 
as evidences of the state of the interior; — that their 
courses are far shorter than those of the long waves; 
that even these transmitted waves are not perfectly sepa- 
rated but are attended sometimes by irregular or adventi- 
tious waves to such an extent as to embarrass their 
interpretation (Walker) ; that they are weakened as they 
penetrate the interior material to greater and greater 
distances until both types die down beyond distinct 
identification; that this takes place before their penetra- 
tions reach the diameter of the earth ; and that therefore 
there is no substantial ground for postulating any special 
separating substance, "undercooled liquid" or otherwise. 
The fatalities of the molten theory. — Among the evi- 
dences from various sources relative to the doctrine of a 
molten earth, this singular effort to fill the interior with 
non-crystalline material similar to "undercooled liquids/ ' 
the effort to divorce rigidity and elasticity from crystal- 
linity by the overstrained affirmation that ' ' It is evident 
that both rigidity and elasticity are totally independent of 
crystaHinity'' (p. 397), the disparagement of the evi- 
dences of elastic rigidity and similar strained efforts, are 
tokens of the desperateness of the endeavor to stem the 
tide of calamity that has been sweeping so mercilessly 
over the molten theory in recent years. First came the 
ancient glaciations in surprisingly low latitudes, those of 
the late Paleozoic in the lead, and then those of the very 
early Paleozoic, followed by those of the Proterozoic, 
which together swept away all climatic evidences of an 
early hot stage. Almost hand in hand with the adverse 
testimony of ice invasions, came the loss of the basal 
credentials of the theory. The early advocates of the 
molten theory were able to point with much show of 
cogency to the great granitic embossments so often found 
lying under the fossiliferous series as direct testimony 
to the light acidic layer of igneous material postulated as 
the outer zone of the molten earth, but the field studies of 
the last few decades have shown that these granitic 
masses are batholitic intrusions into older sedimentary 

Groundwork for Study of Megadiastrophism. 273 

deposits and are no more evidence of an original molten 
state than the great granitic intrusions of Jurassic age 
along our western coast, as recently remarked by Cole- 
man. 16 And then came radioactivity and robbed the 
molten theory of almost its only resource for actuating 
diastropliism, viz. : the cooling of the earth ; while close 
on the heels of this came the seismic waves and their allies 
which are now heaping up irrefragible evidence that, if 
there ever was a molten state, it is now a thing of the past, 
and that the great dynamic problems of the day, diastro- 
pliism, metamorphism, asthenozones, magmatic genera- 
tion, vulcanism, isostasy, paleogeography, and the like, 
must all be worked out on the basis of a solid elastic earth. 
If the molten earth is an issue at all in present day 
problems, apparently it can be only in the sense of an 
issue over the estate left behind by the departed. 

10 A. P. Coleman, Xature, vol. 109, p. 775, June 17, 1922. 

Universitv of Chicago, 
Julv 19, 1922. 

274 J. S. Brown — Relation of Sea Water 

Akt. XXVII. — Relation of Sea Water to Ground Water 
along Coasts; 1 by John S. Bkown. 


Sea water in shallow wells. 
Sea water in deeper, drilled wells. 

Effect of special topographic and geologic conditions. 
Law of equilibrium between fresh and salt water. 
Nature of the contact between fresh and salt water. 
Effects of pumping on salt content. 
Seasonal variations in salt content. 
Influence of tides on ground water. 


Coastal ground-water problems: — The water from 
wells and springs near seacoasts is often salty. On most 
coasts saltiness due to sea-water is limited in shallow 
wells to a narrow zone, usually a few hundred feet in 
width, near the shore, but at some places it extends much 
farther inland. Many wells near the shore encounter 
fresh water near the surface and salt water at greater 
depths. Pumping plants often yield salty water, espe- 
cially after long-continued operation at a high rate. The 
circumstances of contamination by sea water are modified 
greatly by the kind of rocks and their structure and by 
the topography adjacent to the shore, as well as by 
climatic factors. 

Scope of field ivork. — In 1919 the writer made a study 
of the groundwater resources of the New Haven area, 
Connecticut. 2 This area includes a section of the coast 
of Connecticut, lying east and west, of New Haven, which 
will be referred to hereafter as the New Haven coast. 
The length of this section is 30 miles in a straight line, 
but owing to small irregularities, the actual coast line is 
much longer than 30 miles. This work afforded a partic- 
ularly favorable opportunity to observe some of the con- 
ditions under which ground water is contaminated by sea 

1 Published by permission of the Director of the U. S. Geological Survey. 
The illustrations were prepared by the Survey. 

2 Brown, John S. : Ground water in the New Haven area, Connecticut, U. S. 
Geol. Survey Water-Supply Paper [awaiting publication]. 

This work was done in co-operation with the Connecticut State Geological 
and Natural History Survey, under the direction of Herbert E. Gregory. 

to Ground Water along Coasts. 275 

water. 3 Some valuable information also was obtained 
during a few days spent on the coast of Florida in 
April, 1920. 

Previous investigations. — In the United States many 
geologists and engineers have described contamination of 
ground-water by sea water, and a few writers have dis- 
cussed at some length phases of the problem incidental 
to other investigations. However, there has been no 
systematic investigation of the problem or serious attempt 
to assemble and apply the data already published. In 
Europe, the denser settlement and more intensive devel- 
opment of water supplies caused the problem to be recog- 
nized earlier, and in Holland, Belgium, France, and 
Germany many data have been collected and numerous 
papers published on the subject. The present writer has 
used the published material to supplement the informa- 
tion obtained on the New Haven coast and confirm the 
conclusions reached. 

Methods of investigation. — In the course of this investi- 
gation a geologic map of the region was made. Records 
of wells, springs, and pumping plants and details of their 
construction, use, and operation were collected. Topo- 
graphic features were noted carefully. In order to deter- 
mine the degree of contamination by sea water many 
samples of water were analyzed. As in many previous 
investigations, the chloride 4 content of the samples has 
been used as the chief criterion for determining the 
presence or absence of sea water. Chloride is not always 
a reliable indicator of contamination by sea water, for it 
may come either from salt in the rocks or from sewage 
and factory wastes. The rocks of Connecticut, however, 
yield very little salt, and in areas free from polluted 
drainage the ground water contains but little chloride, — 
in fact, 10 or 12 parts per million is usually the maximum 
quantity present from both these sources. 5 In this inves- 
tigation, it was generally possible to tell from the sur- 
roundings whether chloride from sewage or waste was to 

3 A complete report, of which this paper is a summary, entitled ' ' A study 
of coastal ground water, with special reference to Connecticut, " is to appear 
as a Water-Supply Paper of the U. S. Geological Survey. 

4 The term chloride is used here to denote the chloride radicle (CI). 

5 See Jackson, D. D., The normal distribution of chlorine in the natural 
waters of New York and New England: U. S. Geol. Survey Water-Supply 
Paper 144, 1905. 

276 J. S. Brown— Relation of Sea Water 

be suspected, and samples collected where the circum- 
stances seemed to indicate such pollution were rejected. 
A small assay kit was carried in the field for the purpose 
of making approximate determinations of chloride, but 
check samples were sent to the water laboratory of the 
United States Geological Survey for analysis. The lab- 
oratory analyses are given in this paper. 

Sketch of the Neiv Haven coast. — The New Haven coast 
is a "drowned coast " and is consequently very irregular 
in outline. There are large areas of flat tidal marshes 
and in some places many small islands near the shore. 
Portions of the coast have been cut away by waves, and 
other portions have been built up. Spits and bar beaches 
are common features. 

The bed rocks of the New Haven coast consist mainly 
of gneiss and schist. At most places they are concealed 
by glacial drift, which is generally not more than 30 to 
40 feet in thickness, but portions of the coast are bare and 
rocky, and small exposures of bedrock are common over 
much of the region. The drift consists of till and strati- 
fied drift. The till is generally variable in composition 
and texture and is in places somewhat clayey. The strat- 
ified drift has been re-worked by water and is usually 
clean, well-bedded sand or gravel. It is confined mainly 
to the valleys and certain low coastal plains of small 
extent. Besides these formations there is a recent 
deposit of black, slimy mud which covers the bottom of 
New Haven Harbor and the tributary tidal estuaries and 
whose thickness probably does not exceed 25 or 30 feet. 

Occurrence of ground water. — In the bedrocks ground 
water is held mainly in open fractures, such as joints and 
cleavage planes, and the water table is irregular. In the 
stratified drift, which is very porous, the water table is 
in general fairly smooth and stands but little above the 
level of the sea or above adjacent streams. In the till, 
owing to its varying porosity, the water table is irregular, 
but usually water is found within a few feet of the 

Sea water in shallow wells. 

Most of the shallow wells on the New Haven coast 
are either dug or driven and are less than 30 feet in depth. 
More than half of them penetrate stratified drift. The 

to Ground Water along Coasts. 277 

table below shows the number of wells examined in sev- 
eral successive zones parallel to the shore and the number 
and percentage of wells contaminated by sea water in 
each zone. Three degrees of contamination, based on an 
arbitrary scale, are distinguished — " trace" (from 25 to 
100 parts per million of chloride), "moderate" (100 to 
300 parts), and "high" (more than 300 parts). More 
than 300 parts per million of chloride is generally appar- 
ent to the taste. For a few wells where samples could 
not be obtained, positive statements of reliable persons 
have been accepted as evidence of contamination. 

Contamination by sea wafer in shallow wells on the New Haven 


Limits of zone 

No. of 

No. of wells contaminated 

Total No. 


in feet from 



different degi 


of con- 

of con- 





shore line. 







0- 25 







26- 50 


























The above table summarizes very briefly a large mass 
of data from which the conclusion is drawn that on the 
New Haven coast the zone in which the shallow ground 
water is contaminated by percolation or diffusion of salt 
water is very narrow. The greatest distance from the 
shore at which even slight contamination is suspected is 
250 feet. 

Similar conditions were observed at Cedar Keys, Fla., 
where the annual rainfall is very nearly the same as at 
New Haven and where the ground is sandy, comparable 
to the stratified drift of Connecticut. At Cedar Keys 
open pits are often constructed on the beaches of small 
islands to furnish water for stock (fig. 1). The water 
of the pit shown in fig. 1 contained 516 parts per million 
of chloride but was usable for stock. Water in a similar 
pit 120 feet from the reach of normal high tides contained 
only 92 parts per million of chloride. Of six wells in a 
zone between the limits of 125 and 350 feet from the high- 
tide shore line, two were slightly contaminated and in 
another salt water apparently was induced by heavy 


J. 8. Brotvn — Relation of Sea Water 

From the evidence set forth above and that given by 
other writers the conclnsion seems warranted that sea 
water mingles with the shallow ground water over only 
a narrow area near the shore line. Wherever the natural 
conditions are disturbed, however, as through a heavy 

Fig. 1. 

Fig. 1. — Open water hole yielding fresh water on the beach of a small 
island near Cedar Keys, Fla. 

drain on the ground water by pumping, conditions are 
immediately altered and sea water may be drawn great 
distances inland. Also, although data on the matter are 
lacking, it would be logical to suppose that on arid coasts, 
where the rainfall is insufficient to maintain a body of 
ground water, sea water would penetrate a much greater 
distance inland than in humid regions. Undoubtedly, too, 

to Ground Water along Coasts. 279 

the penetration of salt water inland is affected by the 
character and structure of the rocks along the shore. As 
just indicated on sandy parts of the Florida coast fresh 
water saturates the ground almost to the shore line. Yet 
Sanford 6 says that the main Florida Keys, which are 
composed chiefly of limestone very full of seams and 
crevices, "may be underlain by salt water about at tide 
level; this is the condition in islands as wide as 3 or 4 
miles in the Bermudas and also on many Florida Keys." 

Sea water in deeper, drilled wells. 

Most of the deeper wells on the New Haven coast were 
drilled with the percussion type of drill rig and are 6 
inches in diameter. Wells from 100 to 300 feet deep are 
fairly common as sources of domestic supply. A few 
wells of greater depth have been drilled by manufacturing 
concerns in the hope of obtaining industrial supplies, but 
most of these have been failures because the quantity of 
water was insufficient or its quality was unsatisfactory. 
Out of 29 wells examined, 10 were definitely contaminated 
by sea water and 2 others probably were slightly contam- 
inated. Moreover, as wells that yield bad water are 
generally abandoned and forgotten, and their records are 
almost impossible of verification, the ratio is really 
higher. It is not unlikely that half the drilled wells in 
a zone a few hundred feet wide along the shore are more 
or less contaminated. 

Xearly all the drilled wells on the New Haven coast 
penetrate bedrock and draw their water from open joints 
and crevices. The possibility of obtaining water at all 
in the crystalline rocks is difficult of prediction, and the 
likelihood of obtaining salt water in deep wells is doubly 
uncertain. One well very near the shore may penetrate 
crevices connected only with the fresh ground water and 
yield water of good quality; another may penetrate 
crevices connected with the sea and yield only salt water. 
Undoubtedly the danger of obtaining salt water increases 
rapidly with the depth of drilling, and once salt water is 
struck there is very little chance of finding fresh water 
at greater depth. The distance inland at which salt 

Matson, G. C, and Sanford, Samuel : Geology and ground waters of 
Florida, U. S. Geol. Survey Water-Supply Paper 319, p. 261. 

280 J. S. Brown— Relation of Sea Water 

water may be expected in deep wells on the New Haven 
coast probably does not exceed 500 feet at most places, 
because of the fact that a few hundred feet below the 
surface water-bearing fractures in the bedrock are prac- 
tically absent and there is very little circulation of ground 
water. The safe ratio of depth of well to distance from 
the shore is difficult to estimate but is placed at about 
1 to 1 — that is, a well 100 feet from the high-tide shore 
line should not be drilled more than 100 feet below sea 
level if any fresh water has been obtained above that 

On many coasts where geologic conditions are unlike 
those in Connecticut impervious beds seal out salt water 
from underlying strata containing fresh water. This is 
particularly true of coastal plains, such as our Atlantic 
Coastal Plain, where alternating sedimentary beds of 
different kinds dip seaward at low angles. This condi- 
tion is exhibited on a small scale, in New Haven Harbor 
(see beyond). 

Effect of special topographic and geologic conditions. 

Tidal marshes: — Very few wells are sunk within the 
limits of tidal marshes, but from such information as was 
obtained the conclusion seems warranted that most of 
the tidal marshes contain brackish ground water. As a 
rule the salinity of this ground water is much less than 
that of sea water, because the tidal water that overflows 
the marshes is diluted by fresh water from streams as 
well as directly by rainfall. In many places the underly- 
ing ground water is further diluted by ground water from 
adjacent areas. In the more or less stagnant marshes, 
such as those of Milford, Madison, and Clinton, which 
receive but little inflow from streams and are separated 
from the sea only by low, narrow strips of land, the 
ground water probably is very saline. They may be com- 
pared to the sea of Haarlem, in Holland, where salt 
water rises nearly to the surface. (See p. 286.) In nar- 
row tidal marshes along river estuaries the ground water 
is much fresher. 

Bar beaches and spits. — It appears from all the evi- 
dence obtained that the ground water of bar beaches and 
spits is salty. Probably the water table sinks so low at 

to Ground Water along Coasts. 281 

low tide, because of the extreme porosity of the sand, that 
at high tide there is a distinct influx landward of salt 
water. High storm waves sometimes completely overtop 
the bars and spits, thus undoubtedly saturating them with 
salt water. On unusually high and wide spits a small 
amount of usable water might be obtained by shallow 
wells located as far as possible from the sea. However, 
the ground water doubtless becomes exceedingly saline 
at no great depth below the water table on most of these 

Islands. — Conditions on islands do not appear to differ 
materially from those on the mainland except where the 
available intake for ground water is exceedingly small. 
The nature of the land surface and of the rock composing 
the island are factors of importance in influencing the 
character and quantity of the ground-water supply. A 
covering of till or of stratified drift promotes absorption 
of rainfall and equalizes the supply of ground water 
reaching the water table, whereas bare rock surfaces pro- 
mote run-off. It seems from all the available evidence 
that good water, in small quantities, may be obtained even 
on very small islands where there is a cover of till or 
stratified drift. An island having an area of an acre, or 
about 250 feet in diameter, usually will supply enough 
water for an ordinary household. The supply increases 
with the size of the island, though probably in consider- 
ably greater ratio, as the relative losses about the 
perimeter become smaller. 

Effect of bedrock fractures. — Sea water enters wells 
that are dug or drilled in bedrock near the shore through 
fractures. Clapp 7 suggests that in certain contaminated 
wells on a peninsula in Maine the sea water enters at 
the side of the peninsula farthest from the wells, because 
the cleavage of the rocks dips from that side toward the 
wells. From the evidence on the New Haven coast it 
seems reasonably safe to assume that a high degree of 
fracturing in the bedrock on the shore increases the 
chances of sea water entering wells near by, and that the 
trend or inclination of fractures toward a well may also 
be of importance, but only as indicating the greater like- 

7 Clapp, Frederick G. : Underground waters of southern Maine, U. S. Geol. 
Survey Water-Supply Paper 223, p. 67, 1909. 

282 J. 8. Brown — Relation of Sea Water 

likood of the existence of a fracture connecting the well 
with the sea. There is no apparent reason why cleavage 
planes dipping toward a well will admit water much more 
rapidly than a series of connected fractures. It is diffi- 
cult to evaluate the effect of fresh water fed to a well by 
fractures. If the supply is considerable and the water 
is under sufficient pressure sea water will be excluded 
under conditions that otherwise would readily permit it 
to enter the well. 

Artesian conditions in New Haven Harbor. — The 
blanket of mud that forms the most recent deposit on the 
floor of New Haven Harbor and its tributary tidal 
marshes acts as an effective impervious stratum and 

Fig. 2. 

Ifew Haven Ifarbor ■ New Haven N. 
_^-rr ^--W^LULde 

Fig. 2. — Section illustrating artesian conditions in New Haven Harbor. 

excludes the salt water of the harbor and the marshes 
from the sands of the stratified drift beneath. The main 
features of the New Haven artesian system are illustrated 
by figure 2. The sands beneath the mud average 100 to 
200 feet in thickness and are, at least for the most part, 
filled with fresh water. The wells of Sperry & Barnes, 
which are on filled land nearly 1,000 feet south of the 
original high-tide shore line, draw comparatively fresh 
water from the sand beneath the fill and the mud. In a 
test well of Sargent & Co., 660 feet south of the original 
water front, fresh water was obtained beneath the mud 
at a depth of 110 feet. The plants of the New Haven Gas 
Co. and the National Folding Box Co. use wells on the 
filled-in tidal marshes of Mill River and draw fresh water 
from sand beneath the mud. Some of these plants yield 
water very definitely contaminated by sea water, but 
this condition has resulted from excessive pumping. At 
all the places mentioned the water was originally drink- 
able and at present is only slightly brackish. 

to Ground Water along Coasts. 2S3 

Law of equilibrium between fresh and salt water. 

The reason that fresh water can exist in close proximity 
to sea water on small islands composed of freely porous 
material is that it has a lower specific gravity and so rests 
upon the underlying salt water. The slow rate at which 
water percolates through the crevices in rocks and the 
equally slow rate of diffusion of salt through rocks pre- 
vent the rapid mingling of the fresh water with the salt 
water, and the continued increment from rains replaces 
that which is lost. The same conditions exist on main- 
land coasts. As a result, the fresh and salt water reach 
static equilibrium and assume positions which are depend- 
ent on relative specific gravities, the higher column of 
fresh water balancing the heavier sea water. This prin- 
ciple was first recognized by Badon Ghyben, 8 a Dutch 
captain of engineers, as the result of work clone on the 
coast of Holland; it was later enunciated by Herzberg, 
apparently without knowledge of the work of Badon 
Ghyben. Herzberg 9 found in drilling wells on Xorcl- 
erney, one of the East Frisian Islands ofT the German 
coast, that the depth to salt water was roughly a function 
of the height of the water table above mean sea level and 
of the density of the water of the Xorth Sea. Figure 3 
shows the application of his theory. 

Let H be the total thickness of the fresh water. 
Let h be the depth of fresh water below mean sea-level. 
Let t be the height of fresh water above mean sea-level. 
Then H = h + t. 

But the column of fresh water H must be balanced by 
a column of salt water h in order to maintain equilibrium. 
"Wherefore, if g is the specific gravity of sea water (and 
if the specific gravity of fresh ground water is assumed 
to be 1), 

H =z h -{- t = hg, whence 

» = ' 7 


6 Badon Ghyben, W. : Xota in verband met de voorgenomen put boring 
nabij Amsterdam, Tijdschrift van het Konirklijk Instituut van Ingenieurs, 
Instituuts jaar 1888-1889, p. 21, 's-Gravenhage. Referred to by Eugene 

9 Herzberg, Baurat: Die Wasserversorgung einiger Xordseebader, Jour. 
Gasbeleuchtung und Wasserversorgung, Jahrg. 14, Miinchen, 1901. 

284 J. 8. Brown — Relation of Sea Water 

In any case g — 1 will be the difference in specific gravity 
between the fresh and salt water. Herzberg gives the 
specific gravity of the North Sea as 1.027, whence h = 37t. 
Drillings on Norderney and other islands near by gave 
results varying but a few meters from those derived by 
the formula, but it is not clear whether the salt water 
encountered at the expected depth had approximately the 
composition of sea water or was merely too salty for use. 
Herzberg states, further, that the results are sometimes 
greatly modified by the fineness or coarseness of the 
sands, indicating that a factor due to porosity is involved. 
The theory appears to apply particularly to small islands 
and narrow land masses which consist of freely pervious 

Fig. 3. 

Sea"^ tef 

Fig. 3. — Section of the island of Norderney, showing the application of 
Herzberg 's theory. (From Herzberg.) 

material, especially sand, and even to these only to a 
certain extent. However, Herzberg 's paper appeared at 
a time when there was great interest in coastal water 
supplies in Holland and Belgium, particularly at Amster- 
dam, and was widely quoted and much discussed by 
Dutch, Belgian, and French writers. The law of equi- 
librium between salt and fresh water is frequently 
referred to as the "theory of Herzberg.' ' 

Nature of tine -contact between fresh and salt water. 

Along coasts there is a continual diffusion of salts from 
the sea water into the fresh ground water of the land and 
an actual movement of fresh ground water seaward. 
There must exist a transition zone in which equilibrium 
is established between these opposing forces. The loca- 
tion and nature of this zone are determined in part by 

to Ground Water along Coasts. 


several other factors, including the tides, (see p. 280), the 
permeability and structure of the rocks, the amount of 
fresh water supplied by rainfall, the density of the sea 
water, and perhaps the prevailing temperature. As has 
been shown by Herzberg and others, the depth of the 
contact below mean sea-level is, in uniform material, a 
function of the height of the water table above sea-level. 
The simplest case that can be assumed for considera- 
tion is that of a small island composed of uniformly por- 
ous sand. The probable movements of ground water and 

Fig. 4. 

Fig. 4. — Ideal diagram showing movements of ground water and grada- 
tion into sea water along shores composed of uniformly porous sand. A, On 
a small island: B, on a mainland shore. 

the gradation from salt to fresh water on such an ideal 
island are illustrated in figure 4 A, which shows the 
approximate form of curve given by Herzberg for the 
island of Norderney. Imbeaux 10 states that the form 
of curve which he obtained experimentally closely 

10 Imbeaux, E. : Les nappes aquif eres au bord de la mer ; salure de leurs 
eaux, Bull, des seances de la Societe des sciences de Nancy, ser. 3, t. 6, pp. 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 22. 

-October, 1922. 

286 J. S. Brown — Relation of Sea Water 

approached a sinusoid. It appears that as the permea- 
bility of the rocks increases the curve flattens. Thus 
Sanf ord has described the body of fresh water above the 
salt water in the fractured limestones of the Florida 
Keys, as a "thin sheet" (see p. 279), and Lindgren 11 has 
found practically the same condition in the very porous 
basalts of Molokai. This is a very natural result, for 
the greater freedom of circulation permits the water table 
to sink lower, and the necessity for equilibrium then 
compels the salt water to rise higher, so that the two 
approach each other. On coasts composed of porous 
sand the conditions are probably somewhat as shown in 
figure 4 B. The general zone of contact slopes downward 
beneath the coast until it is interrupted by impervious 

In Connecticut this invasion of the land by sea water 
terminates within a few hundred feet of the shore, owing 
to the fact that the underlying crystalline rocks become 
practically impervious within a few hundred feet of the 
surface. Pennink, 12 however, has demonstrated most 
convincingly that salt water extends for several miles 
inland beneath the coast of Holland and similar condi- 
tions probably exist at other places. The west coast of 
Holland consists of a belt of dunes about 5 miles wide 
and at its crest 10 to 20 meters above sea level, which 
separates the sea from a belt of polders. 13 The polders 
were created by draining the sea of Haarlem about 1850 
and lie from 4 to 6 meters below sea level. In seeking 
to supplement the Amsterdam water supply a line of 
test wells was driven across this belt of dunes. Chloride 
determinations were made on the ground water at many 
different depths in these wells, and a cross section show- 
ing the actual contact between fresh water and under- 
lying salt water was determined. (Figure 5.) 14 Certain 

11 Lindgren, Waldemar : The water resources of Molokai, U. S. Geol. Sur- 
vey Water-Supply Paper 77, pp. 26-47. 1903. 

12 Pennink, J. M. K. : De " prise d 'eau ' ' der Amsterdamsche duin water- 
leiding, Tijdschrift van het Koninglijk Instituut van Tngenieurs, pp. 183- 
238, The Hague, 1904. 

13 In Holland and Belgium the term polder is applied to a tract of marshy 
land, lower than the sea, which has been diked and reclaimed to cultivation. 
— Standard Dictionary. 

14 The zero of Amsterdam shown in Pig. 5 is approximately 0.2 meter 
above mean sea level. 

to Ground Water along Coasts. 



o o o 
— pj n 

I I I 

O O o o o 

in \o t~~ co <?> 

I I I I I 

288 J. 8. Brown — Relation of Sea Water 

irregularities in the distribution of salt and fresh water 
are easily accounted for by the local geology, but it is 
clear (1) that salt water actually underlies the land at a 
depth of 100 to 200 meters below sea level over a belt 
several miles in width ; (2) that the depth to salt water 
is greatest where the land and the water table are highest ; 
(3) that salt water rises under areas of low ground, or 
rather under areas having a depressed water table, such 
as the polder of the old sea of Haarlem. The general 
line of contact between salt and fresh water is very 
regular, however, and the zone of gradation or diffusion 
is surprisingly thin. 

Laboratory experiments by D'Andrimont, 15 a Belgian, 
also gave results very much like thos«e from Holland. 

Effects of pumping on salt content. 

In the work near New Haven 23 pumping plants within 
500 feet of the shore were examined. The water at ten 
of these plants was highly contaminated with sea water, 
and that at three or four others was contaminated 
slightly. The greatest distance from the shore at which 
high contamination was noted was 400 feet. In general, 
the plants that supplied the larger amounts of water were 
more highly contaminated. 

All the evidence regarding the effect of pumping tends 
to one conclusion — that if more water is removed than 
the contributory area supplies, the wells become salty. 
This statement is as true for domestic wells drawing from 
a small intake area as for pumping plants supplied with 
the underflow of a large drainage basin. Thus the 
Beattie dug well, which probably drains about 1 acre of 
sand (stratified drift) at the end of a small island, origi- 
nally yielded good water, but the demands of one family, 
when increased by the installation of a household water 
system, were sufficient to overtax the well and make the 
water salty. 

The best data regarding the limiting rate at which 
water may be pumped from a given contributory area 

15 D 'Andrimont, Bene : Note preliminaire sur une novelle methode pour 
etudier experimentalement 1 'allure des nappes aquif eres dans les terrains per- 
meable en petit, Annales de la Societe geologique de Belgique, t. 32, 
Memoires, pp. 115-120, Liege, 1906. 

to Ground Water along Coasts. 28$ 

are furnished by the plant of the American Steel & Wire 
Co. and the connected plant of the Penn. Seaboard Steel 
Corporation at New Haven. When these plants were 
engaged on war work during 1918, and the demand for 
water was excessive, the Penn. Seaboard Steel Corpora- 
tion drew from its wells about 330,000 gallons a day or 
10 million gallons per month. The water, which at first 
was sweet and drinkable, became very salty, evidently 
from salt water drawn in from the harbor. On the small 
plain upon which these plants are located not more than 
a quarter of a square mile serves as a catchment area 
for ground water. It appears, therefore, that the rate 
of pumping greatly exceeded the ability of the catchment 
area to supply water. The rate of 330,000 gallons a day 
for a quarter of a square mile is somewhat more than 
a million gallons a day to the square mile, a figure which 
is considerably above the safe draft upon a drainage basin 
near the sea under conditions such as prevail in Connecti- 
cut. One million gallons a day to the square mile is 
20.8 inches a year, or about 45 per cent of the total annual 
rainfall in the New Haven region. 

The effects of pumping on a large scale near the sea 
have been carefully studied by the Board of Water 
Supply for the City of New York. 16 For an intake area 
of 157 square miles on Long Island, where the climate is 
about the same as in Connecticut and where the surface 
formation is glacial drift, Spear estimates the safe yield 
at 77 million gallons a day, or very nearly half a million 
gallons a day to the square mile. This figure is a much 
closer approximation to the safe capacity of ground water 
reservoirs under such conditions. 

The limit to which wells can safely be pumped or the 
depth to which the water table can safely be lowered can 
never be exactly determined, except by actual tests. The 
statement made by many writers, that if the water table 
is lowered below mean sea level salt water is induced to 
flow inland, is only partly true. An appreciable gradient, 
several feet to the mile, is necessary to permit ground 
water to flow through the rocks. Therefore, where salt 
water extends only a short distance beneath the coast, a 

16 Spear, W. E. : An additional supply of water for the city of New York 
from Long Island sources, New York Board of Water Supply, vol. 1, pp. 
144-159. 1912. 


J. 8. Brown — Relation of Sea Water 

well a mile or two inland could be pumped considerably 
below mean sea level before the salt water would have 
head enough to enter it. 

Seasonal variations in salt content. 

Certain wells that are contaminated with sea water 
exhibit a pronounced variation in salinity. The Peter 
Beattie dug* well is on a small stratified drift plain about 
200 feet in diameter at the north end of Narrows Island, 

Fig. 6. 

Fig. 6. — Northern end of Narrows Island, Guilford, Connecticut. A well 
near the house yielded water with a chloride content that ranged from 83 
to 609 parts per million in one year. 

(fig*. 6). The well is 12 feet deep, in sand and gravel, and 
the water level stands at about mean sea level. The 
results of chloride determinations made from samples 
collected at different dates, covering almost a year, are 
given in f^g. 7 and are compared with the rainfall and 
temperature records. The data show that there was a 
very great increase of saltiness in the water during June, 
July, and August, 1919. 

A factor that should not be overlooked in the study of 
the salinity curve of this well is the quantity of water 

to Ground Water along Coasts. 


pumped for use. In summer the quantity of water used 
for domestic purposes might be expected to increase 
somewhat. However, as the salinity of the water became 

Fig. 7. 



From U.S. Weather Bureau records at New f/<3i/e/7 

Fig. 7. — Eelation of salinity to temperature and rainfall in Peter Beattie 

greater, its use for many domestic purposes was dis- 
continued, and water was carried from a neighboring- 
well. The factor of use therefore does not appear to 
have had any appreciable influence on the degree of 

292 J. S. Brown — Relation of Sea Water 

The variation shows the influence of variations in pre- 
cipitation and apparently also in temperature. The 
extremely high chloride in June and the first part of July 
represents the effect of a month of very light rainfall, 
accompanied by rising temperature. The steady de- 
crease of chloride from July 12 to July 26 is the result 
of heavy, well-distributed rainfall, and the increase in the 
week of July 26 to August 2 coincides with a week of 
light rainfall. The great decrease in chloride between 
August 2 and September 6 is. evidently the result of heavy 
rains, particularly on September 2 and 3, when 3.72 inches 
fell. The increase noted in the following week, Septem- 
ber 6-13, apparently represents the effect of gradual 
rediffusion of salt after the flood of fresh ground water 
had subsided. Undoubtedly careful tests at more fre- 
quent intervals would show many irregularities of the 
curve due chiefly to the variable precipitation, but the 
general form would agree more or less closely with that 
of a yearly temperature curve, although the spring rise 
would lag considerably behind the rise in the temperature 
curve, because the frost is slow to leave the ground in 
spring — that is, the ground warms more slowly than 
the air. 

The influence of temperature appears in different ways. 
Higher temperature increases evaporation. It also 
greatly increases the rate of percolation, so that ground 
water seeps out to sea more freely and salt water diffuses 
landward faster. Slichter 17 states that a change of tem- 
perature from 32° to 75° F. will practically double the 
power of a soil to transmit water, and that a change from 
50° to 60° will increase the rate of flow 16 per cent. The 
influence of humidity is also a factor, and the rate of 
transpiration by plants is important. The island on 
which the well is situated supports a grove of trees and 
a grassy lawn, which transpire much of the precipitation 
in the summer season. The net result, as indicated, is a 
remarkable increase of salinity in summer. 

Observations on two drilled wells in bedrock, although 
not so complete as those on the Beattie well, strengthen 
the evidence that certain contaminated wells are most 

17 Slichter, Chas. S. : Field measurements of the rate of movement of 
underground waters, U. S. Geol. Survey Water-Supply Paper, 140, p. 13, 

to Ground Water along Coasts. 293 

salty in summer. The drilled well of Robert Mitchell, at 
Sachem Head, Guilford, is on a narrow point 50 feet 
from high tide on two sides. It is about 50 feet deep. 
The well was drilled for the supply of a summer cottage, 
but the water became so salty that the well was entirely 
abandoned in 1919. The following quantities of chloride, 
in parts per million, were found in four samples examined 
in 1919: May 19, 49; June 16, 94; Aug. 4, 50; 
Nov. 28, 31. 

The water of the Anderson drilled well on Mulberry 
Point, Guilford, becomes salty whenever it is used. 
Nevertheless it is used in summer, as it is the only supply 
available. The quantity of water used does not vary 
greatly from day to day, and the results given below 
represent mainly natural seasonal variations in salinity. 
Some analyses made in 1919 showed the following quanti- 
ties of chloride, in parts per million : June 16, 37 ; Julv 
12, 290; July 17, 516; July 20, 446; Aug. 4, 355. 

The results indicate a seasonal variation in salinity 
much like that in the Beattie well, though the great 
increase between June 16 and July 12 may be largely due 
to the drain upon the well. The mounting salinity early 
in July shows the effect of a sharp drought and subse- 
quent decreases closely follow rains on July 16 and 
August 1. 

.Such remarkable fluctuations in salinity in ground 
water must be confined to small islands and isolated 
points Avhere the total quantity of fresh ground water is 
very small and where the increment from rains reaches 
the sea quickly. It is quite possible, however, that at 
some places on mainland coasts seasonal fluctuations have 
an appreciable effect on the zone between salt and fresh 
ground water in which contamination by sea water occurs 
and may cause this zone to progress slightly farther 
inland in summer. The evident influence of temperature 
also suggests that in warm climates contamination is 
relatively greater than in cooler regions with a like 
amount of rainfall. 

Influence of tides on ground water. 

The water surface in wells near the sea is sometimes 
affected by tides, but the writer's observations indicate 

294 J. 8. Brown — Sea and Ground Water. 

that fluctuations of this kind in sandy material are less 
than a few tenths of a foot. In fractured rocks, however, 
fluctuations of several feet may occur. The best data 
on tidal fluctuations in wells are given by Veatch, 18 who 
says thart they are caused in three ways — "by transmis- 
sion of pressure through open cavities or passageways 
affording a free communication between the wells and 
the ocean ; by a checking of the rate of discharge of the 
normal ground-water flow through porous beds freely 
connecting with the ocean ; and by a deformation of the 
strata due to the alternating loading and unloading of 
the tides.' ' The last method accounts for the rhythmic 
rise and fall of flowing wells near the sea. 

Experiments indicate that fluctuations in salinity in the 
water from wells can only very rarely be correlated 
directly with the rise and fall of the tides. The tidal 
period is too short for its immediate effect upon the salt 
content of ground water to be propagated in a measurable 
amount much beyond the limits of the low-tide and high- 
tide shore line. However, when wells near the sea are 
pumped so heavily that the water table is kept below sea 
level for periods of several days or more and salt water 
is thus enabled to flow toward the wells, the rate of this 
flow increases at high tide due to the increased head of 
salt water and a fluctuation in salinity having the perio- 
dicity of the tides is established. 19 

Undoubtedly tides are very important in maintaining 
the zone of brackish water that occurs along the coast. 
They act partly through direct infiltration of salt water 
at high tide (Ag. 4), and partly through oscillations in 
the contact zone due to variations in hydrostatic pressure. 
These oscillations promote intermixture of fresh and salt 

18 Veatch, A. C. : Fluctuations of the water level in wells, with especial 
reference to Long Island, New York, U. S. Geol. Survey Water-Supply 
Paper 155, p. 69. 1906. 

19 Burr, W. H., Hering, Eandolph, and Freeman, J. E., Report of the com- 
mission on additional water supply for the city of New York, pp. 406-423, 
New York, Martin B. Brown Co., 1904. 

A. Wandke — Study of Cape Neddick Gabbro. 295 

Art. XXVIII. — A Petrologic Study of the Cape Neddick 
Gabbro; by Alfred Wandke. 

Location and Associated Rocks. 

The Cape Neddick gabbro, an elliptical stock, elongated 
northwest, measuring three fourth's by half a mile, forms 
the major portion of Cape Neddick, York County, Maine. 
The stock invades the Kittery quartzite, a fine-grained 
thinly bedded and vari-colored formation of steeply 
inclined sediments which strike about 45 degrees north- 
east, dip 60 degrees northwest and are presumably of 
Upper Carboniferous 1 age. The intrusive may then be 
dated tentatively as post Upper Carboniferous. 

In addition to this intrusive of gabbro the shoreline 
portion of this part of Southwestern Maine offers a vast 
assemblage of dike rocks which not only illustrate most 
of the features associated with dike intrusion but also 
show contrasted types as diabases, diorites, quartz dio- 
rites, granite porphyries, tinguaites, camptonites and 
aplites. Within half a mile of the stock is a batholith of 
alkaline granite, but one of the rock varieties that make 
up the composite mass known as the Agamenticus com- 
plex. This complex shows in addition to a normal biotite 
granite, alkaline rocks such as alkaline granite, syenite 
and nordmarkite. 

The stock rising 45 feet above tide is well exposed. A 
small portion of the landward half of its surface is drift- 
covered, whereas the seaward half is almost swept clean 
of debris. Excellent exposures are thus to be had not 
only of the contacts permitting the contact phenomena to 
be studied in detail, but also of the igneous mass itself 
which is seen to be -composed of contrasting rock types 
in rather complex relationships. 

Method of Emplacement. 

The debris-free contact zone, which in places is 
exposed two hundred feet away from the intrusive, 
enables one to study the mechanism of intrusion as well 
as the details of contact metamorphism. As is shown in 
fig. 1 the stock is clearly cross cutting. Although the out- 

1 U. S. Geol. Survey, Prof. Paper 108, p. 165, 1917. 

296 A. Wandke — Study of Cape Neddich Gabbro, 

Fig. I. 


I*** I (to. bbr o 

W1 Alk<l)ine Granite 

E22 KitieryQuart^ie { 

Fig. 1. — Map showing the geology of Cape Neddick and vicinity. 
Scale 2 inches^i mile. 

A. Wandke — Study of Cape XeddicJc Gabbro. 297 

line of body is oval there is a distinct flattening of the 
southwestern margin along which the contact phenomena 
are also less intense. Along the northeastern contact a 
marine terrace, well exposed at low tide, shows in part 
how the emplacement of the gabbro was effected. 

As has been mentioned, the qnartzite formation strikes 
about 45 degrees NE. At a distance of two hundred feet 
from the contact this strike remains undisturbed. At 
about one-hundred and twenty-five feet from the contact 
the strike of the sediments becomes slightly flexed. 
Within the next twenty-five feet this flexing increases in 
amount and passes into gentle folding the folds becoming 
steeper the nearer one approaches the contact until at 
seventy-five feet therefrom they pass into overturned 
folds which in turn are broken by overthurst faults. 
About fifty feet from the contact the sediments that else- 
where in this region form an easily recognized lithologic 
unit having a remarkable persistency of strike and dip 
lose their identity and become a jumble of blocks of 
quartzite cemented by gabbro. "Within five feet of the 
contact the sedimentary nature of some of these blocks 
is almost obliterated. The actual contact because of the 
marked alteration of the sediments is established with 

To emplace this body a force would have been required 
great enough to bodily thrust the enclosing rocks 
asunder, at the same time folding them and producing a 
shatter breccia. But such a lateral thrusting could not 
alone account for the full width of the stock. The stock 
is half a mile across whereas the total visible movement 
in the contact zone about the intrusive accounts for but a 
few feet of this width. In the contact zone the sediments 
are brecciated and blocks of quartzite are frozen in the 
gabbro. This would at once suggest that magmatic 
stoping has been operative, but stoping hardly accounts 
for the width of this intrusive for the expansion in 
volume of the last 500 feet of stoped sediments would 
have entirely filled the space now occupied by the gabbro. 
It seems, therefore, that stoping may have played but a 
very minor part in the emplacement. 

The emplacement of this body may be accounted for 
if one considers that the intrusive moved upward in a 
zone of tension. A rupture within this zone would par- 

298 A. Wandhe — Study of Cape Neddich Gabbro. 

tially have satisfied the stresses involved and along this 
break magma nnder a high hydrostatic head may have 
risen. If the tensional stresses within the sediments 
were not entirely satisfied by the ruptnre it seems 
probable that a lateral creep of walls may have taken 
place aided by the hydrostatic pressure of the ascending 
magma. If the hydrostatic pressure of the magma and 
the resistance to movement of the solid walls were still 
unequally balanced the enclosing walls may have been 
thrust apart until equilibrium ensued. The field rela- 
tions indicate that such a thrusting apart of the walls 
actually took place. After the intrusive had approxi- 
mately attained its present size it may be that the contact 
which in the first stage had been left jagged was now 
smoothed off. Stoping and local assimilation, processes 
indicated by included blocks and marginal hybrids, evi- 
dently were responsible for this final smoothing of the 


In speaking of this intrusive rock as a gabbro atten- 
tion is called merely to the dominant marginal phase. 
This stock in reality shows four distinct phases: (1) a 
marginal or contact phase characterized by hypersthene 
and approximating a quartz-norite in composition; (2) 
gabbro; (3) and (4) two contrasting central phases, the 
one characterized by ferromagnesian minerals and sug- 
gesting cortlandite, the other characterized by plagioclase 
feldspar and best described as an anorthosite. An opti- 
cal analysis by the Rosiwal method gives the following 
approximate mineral composition of each of these four 
phases : 

Contact Gabbro Central 

Olivine 17.89 1.50 6.76 

Hornblende 10.01 21.42 .68 

Pyroxene 23.53 25.00 8.23 9.43 

Biotite 3.77 3.26 4.76 1.21 

Plagioclase 45.50 45.60 51.85 87.10 

Alkaline Feldspar . . . 1.17 .16 

Magnetite 3.62 13.82 6.79 1.55 

Apatite 2.97 

Hypersthene 2.33 .... .... .... 

100.78 99.19 99.81 100.13 

A. Wandke— Study of Cape Neddick Gabbro. 299 

Contact phase. — This phase forms a peripheral zone of 
tough hard rock that rarely exceeds five feet in width. 
Inclusions of quartzite in various stages of assimilation 
are locally abundant. In the hand specimen the rock has 
a dark gray color, is fine-grained although the crystals 
of olivine that are abundantly sprinkled throughout the 
specimen may attain a maximum diameter of three milli- 
meters. Pyroxene and plagioclase can also readily be 
identified. The presence of rather abundant apatite in 
this contact phase evidently indicates that volatile com- 
ponents played an active part in assimilating the engulfed 
quartzite, the interstitial quartz and micropegmatite 
being corroborative evidence of such action. Judged 
from its mineral composition the rock is a hybrid and 
resulted from the assimilation of quartzite by gabbro. 

Gabbro. — The gabbro forms an interesting phase of 
this intrusive. In part it is developed as a normal 
course-grained dark colored igneous rock, but for a width 
of twenty-five to fifty feet from the contact phase it is 
strikingly banded. These bands rarely over three inches 
wide follow the contact, stand vertically and are due to 
slight fluctuations in the relative amounts of feldspar and 
ferromagnesian minerals. As distance away from the 
contact is gained this banding becomes vague and disap- 
pears, the rock then having the appearance of normal 
coarse-grained gabbro. In the bands the order of crys- 
tallization of the essential constituents seems to have 
been olivine, magnetite, plagioclase, pyroxene, horn- 
blende, and biotite. The olivine, never abundant, occurs 
in grains with an anhedral outline that indicates resorp- 
tion. The magnetite is irregular in outline, is commonly 
enclosed in the feldspar sometimes occurring as dust-like 
particles, and also occurs in the pyroxene surrounded 
by a corona of biotite. Plagioclase feldspar (Ab 2 An 3 ), 
the most abundant single constituent, was one of the first 
minerals to form and its period of crystallization over- 
lapped that of the pyroxene, hornblende and biotite. It 
is frequently euhedral in outline, the ferromagnesian 
minerals being moulded about the well formed crystals. 
The pyroxene characteristically alters to hornblende and 
biotite, a change that appears to have been intermittently 
favored. It may well be that since they follow the pyrox- 
ene both hornblende and biotite, minerals containing 
hydrogen and artificially produced only in the wet way, 
indicate the presence of mineralizers. 

300 A. Wandke— Study of Cape Neddick Gabbro. 

The banding in igneous rocks has received considerable 
attention from petrographers and many examples have 
been described in detail. Various theories have been 
advanced to explain these banded rocks and it seems quite 
certain that different processes have been operative in 
different localities to produce a somewhat similar result. 
The entire subject has been well reviewed by Dr. F. F. 
Grout 2 who gives the following ways in which these 
banded rocks may have been produced: 

1. Partial assimiliation of inclusions forming schlieren. 

2. Lit par lit, or fluidal gneiss. 

3. Deformation during solidification. 

5. Streaked differentiation with reference to rhythmic cooling 

or intrusive action. 

6. Successive intrusions. 

(a) Cooling separately and successively. 

(b) Cooling later, all together. 

7. Heterogeneous intrusion. 

8. Convection during crystallization differentiation. 

At Cape Neddick where the bands are rarely over three 
inches wide, stand vertically, have a granitic texture, 
show but slight variations and can be followed for 
hundreds of feet about the contact, most of the above 
listed explanations such as 1, 2, 3, 6, and 7, can obviously 
be dismissed as non applicable. The facts needing expla- 
nation are the localization of the banded rock near and 
parallel to the contact; a vertical banding; alternation 
between femic and salic rock types; the alteration of 
pyroxene to hornblende and biotite with partial absorp- 
tion of magnetite; the passage of banded gabbro into 
unhanded gabbro toward the center. 

The narrow width of the bands and the passage from 
banded rock into unbanded rock would seem to rule out 
the possibility of successive intrusions. There remain 
thus as the favored hypotheses either convection during 
crystallization differentiation or rhythmic 3 cooling. 
Were it a question of explaining a banding that showed 
gradations from basic to more acidic rock types or vice 
versa then convection might have played an important 
part. A study shows that the bands differ only in the 

2 Journal of Geology, vol. 26, p. 439, 1919. 

3 N. V. Ussing : Geology of the Country Around Julianehaab, Greenland, 
p. 361, 1911. 

A. Wandke — Study of Cape Neddick Gabbro. 301 

relative amounts of light and dark constituents, any two 
adjacent bands throughout the banded area being seem- 
ingly like any other two adjacent bands and the two 
selected bands differing perhaps but little in average 
composition from the unbanded rock into which the 
banded rock grades. Some process, acting intermit- 
tently and rhythmically, for the bands are approxi- 
mately of the same width in any random two-foot zone, 
would best account for the field relationships. It seems 
to the writer that the pulsatory escape of mineralizer may 
best be appealed to as the mechanism responsible for the 
banding. The steeply inclined thinly bedded sediments 
would have furnished excellent avenues of egress to the 
volatile components. These sediments, moreover, for 
over one-hundred feet distant from the contact are not 
only recrystallized but show evidence of the addition of 
new constituents. The abundant apatite in the contact 
phase as well as the marked alteration of pyroxene in 
the banded phase to hornblende and biotite also indicate 
the presence of mineralizers. The escape of the volatile 
components would at once upset pressure, one of the 
factors of temporary equilibrium. This upset in pres- 
sure would manifest itself not in a local change in equili- 
brium but in a change that would at once be transmitted 
throughout the entire magma. If now the body be 
thought of as crystallizing in the normal way, from the 
margin inward, then the change in equilibrium might 
make itself most apparent in this marginal zone, a banded 
rock being the result of recurring upsets in equilibrium. 

Central phases. — The central portion of this stock is 
occupied by two -phases showing as marked differences in 
composition as do any two adjacent bands of the pre- 
viously described banded phase of the gabbro. The 
surface of this central portion, practically clear of debris, 
could be studied in detail and the phases mapped without 
exaggeration. As shown by the map the relations are 
almost such as to suggest that these two contrasted 
phases had differentiated and then the magma in the 
stock been given a swirl with a gigantic mixing spoon. 
The essential minerals of the dark phase are olivine, 
magnetite, pyroxene, pyrrhotite, hornblende, biotite and 
labradorite. The hornblende and biotite both replace the 
pyroxene. The biotite in addition forms about and 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 22. — October, 1922. 

302 A. Wandhe — Study of Cape Neddick Gabbro. 

resorbs the magnetite. The essential minerals of the 
light phase are labradorite, biotite, pyroxene, magnetite 
and hornblende, the dark minerals being accessory to the 
feldspar with biotite the most abundant ferromagnesian 
mineral. The striking feature of the thin sections of 
these rock types is the anhedral fragment-like habit of the 
silicate minerals. It is as though movement had taken 
place in this mass just before final complete solidifica- 
tion was accomplished. It is possible, therefore, that in 
the later stages of crystallization of this stock differen- 
tiation took place largely by the settling of crystals 4 
the tendency being perhaps for the ferromagnesian and 
metallic minerals to sink leaving in the upper portion of 
the stock an accumulation of crystals consisting for the 
most part of feldspar. Just before complete consolida- 
tion was effected movement in this crystal mush may 
have taken place, the relationships as now seen being the 

Secondary alteration. — Eecent road-cuts on the Cape 
made available fresh specimens of the various rock 
phases that to the eye appear unaltered. The micro- 
scope, however, shows that the plagioclase, biotite, and 
olivine have undergone slight alteration. The plagio- 
clase alters to sericite and calcite, the biotite to chlorite 
and calcite, the olivine to serpentine. Kaolinization of 
the feldspar cannot be detected nor is limonite a product 
of the alteration affecting either the biotite or the olivine. 
The changes are, therefore, ascribed to post consolida- 
tion solutions of the general nature of those which 
usually accompany the ores of mineral deposits. The 
slight amount of change produced in the rock minerals 
would indicate that post consolidation hypogene solutions 
were but sparingly present. 

Contact actions. — The shearing and mashing of the 
enclosing sediments have already been described. The 
contact metamorphism of the quartzites is also well 
shown. Away from any intrusive the quartzites are 
seen to consist of quartz, with minor amounts of calcite, 
biotite, chlorite, plagioclase, orthoclase, and rutile. Two 
suites of specimens were studied from the contact zone, 
the one consisting of three specimens taken one-hundred- 

4 N. L. Bowen, Journal of Geology, vol. 27, No. 6, 1919. 

A. Wandke— Study of Cape Xeddick Gabbro. 303 

fifty, fifty and one foot from the contact; the other of 
three specimens taken fixe, three and one foot from the 
contact. In each of the thin sections made from these 
specimens recrystallization of the original minerals and 
the addition of material are indicated by the development 
of feldspar, diopside, biptite, epidote, brown hornblende, 
calcite, magnetite, and pyrrhotite. In the specimens 
taken one-hnndred-fifty feet from the contact the diopside 
and other minerals as listed above are not abundantly 
developed. But as the contact is approached both feld- 
spar and diopside increase greatly in amount. The feld- 
spar which at a distance consisted of orthoclase and albite 
is decidedly more basic and of larger grain near the 

The amount and kind of alteration of the quartzite also 
shows slight variations. Thus a specimen taken five feet 
from the southeast contact is characterized by an abun- 
dance of andesine feldspar which is replacing the quart- 
zite. The development of the plagioclase was evidently 
so rapid that portions of the original quartzite were 
incorporated within the plagioclase, the result being 
grains of feldspar loaded down with inclusions. At the 
contact diopside is a dominant mineral and quartz, 
although present, occurs in decidedly less abundance than 
one-hundr eel-fifty feet from the contact. A specimen of 
quartzite from within the hybrid contact zone of the 
gabbro was also sectioned. The section shows a remark- 
able development of diopside which suggests that this 
mineral was soaked up by the quartzite as a sponge 
soaks up water. The presence, moreover, of hornblende, 
biotite, apatite, magnetite and pyrrhotite would indicate 
that the rock which started as a slightly argillaceous 
quartzite now has the composition of a quartz cliorite. 
Thus there is a tendency for the contact phase of the 
sediments by the addition of material from the gabbro to 
approach a gabbro as far as mineral composition is con- 
cerned and to resemble the phase of the gabbro which 
has been altered by the assimilation of quartzite. 

304 A. Wand/he — Study of Cape Neddick Gabbro. 


The striking features of this Cape Neddick locality 

1. The intrusive apparently occupies a zone of tension 
the emplacement having been effected by lateral creep 
and by bodily thrusting apart the enclosing walls. 

2. Stoping and marginal assimilation are both indi- 
cated but seem to have been rather insignificant processes 
during the emplacement of the gabbro. 

3. The gabbro shows a banded phase that is best 
explained by the rhythmical reduction of pressure within 
the gabbro by the escape of volatile components during 

4. Differentiation by crystal settling seems indicated 
and the basic phase seems later to have intruded the acid 

5. The contact actions indicate the tendency of the 
quartzite near the contact to approach a quartz diorite in 
composition, the gabbro in turn becoming siliceous by 
absorbing quartzite. 

W. H. Dall — Fossils of the Olympic Peninsula. 305 

Art. XXIX. — Fossils of the Olympic Peninsula; by 
William Healey Dall. 

In the Transactions of the Kansas Academy of Science, 
volume 22, pages 131 to 238, Mr. Albert B. Reagan 
described the Olympic Peninsula of Washington, dis- 
cussed its geology and described and figured a number of 
fossils from the Clallam and Quillayute formations. 2 

At that time, paleontological literature was rarely 
accessible in that region, and there had been no modern 
discussion of the late Tertiary fossils of the Oregon- 
Washington area. It is not surprising therefore, that 
identification of the species collected was difficult, and 
sometimes failed in accuracy. 

Fortunately Mr. Reagan turned over his types to the 
IT. S. Geological Survey, which transmitted them to the 
National Museum where they are permanently preserved. 
In working up the fossil faunas of the later Tertiaries of 
the Xorthwest Coast, it became desirable to revise Mr. 
Reagan's list and supply the modern designations, com- 
paring his material with the types of Conrad and others 
from the same general group. Mr. Reagan's species were 
figured somewhat crudely in his memoir, but the types 
were readily correlated with the figures, only two or three 
being absent from the original series. 

His lists were distributed in two parts, one part 
specially devoted to the species of the " Clallam" forma- 
tion, and the other to those of the "Quillayute" 
formation, so called. 

So far as the material in hand goes, the two faunas do 
not seem to differ materially, the species being mixed at 
Coos Bay, but for convenience in referring to Mr. 
Reagan's memoir they will be separately considered here. 

Fossils of the Clallam Formation. 

Solemya ventricosa Conrad. 

Solemya ventricosa Conrad, Report on the Geolog}^ of the U. S. 
Exploring Expedition p. 723, pi. 17, figs. 7, 8, 1849; Reagan, 
p. 171, pi. 1, fig. 1, 1909. 

1 Published by permission of the Director of the U. S. Geological Survey. 

2 Prof essional Paper 59 of the U. S. Geol. Survey was issued April 2, 
1909; Mr. Reagan's memoir, Nov. 24, 1909. 

306 W. H. Ball — Fossils of the Olympic Peninsula. 

Both Conrad's type and that of Mr. Reagan are poorly pre- 
served, but there seems no reasonable doubt of their specific 

Nucula (Acila) conradi Meek. 

Nucula divaricata Conrad, this Journal, vol. 5, p. 432, fig. 1, 

1848; not of Hinds, 1843. 
Nucula conradi Meek, Checklist Miocene foss. of N. Am., Nov. 

Nucula (Acila) castrensis Reagan, p. 174, pi. 1, fig. 2, 1909 ; not 

of Hinds. 
This species has finer sculpture than N. castrensis the recent 
species, and is not so abruptly truncate behind. 

Nucula (Acila) gettysburgensis Reagan. 
Nucula (Acila) gettysburgensis Reagan, p. 175, pi. 1, fig. 3, 1909. 
This is a fine large species closely related to the recent Japanese 
N. mirabilis Adams and Reeve. 

Yoldia reagani n. nom. 

Yoldia impressa (Conrad) Reagan, p. 177, pi. 1, fig. 4, 1909 ; not 
of Conrad. 
A comparison with Conrad's type shows that the latter is 
vertically more attenuated toward both extremities and more 
compressed laterally, being undoubtedly distinct from the present 
shell, which is closely related to Yoldia beringiana and Y. mon- 
teryensis Dall, of the recent fauna. Both of them are of greater 
height from base to umbones, and more inflated than the fossil. 

Phacoides acutilineatus Conrad. 

Lucina acutilineata Conrad, Rep. Geol. U. S. Expl. Exp., p. 725, 

pi. 18, figs. 2, 2a-b, 1849. 
Phacoides acutilineata Reagan, p. 179, pi. 1, fig. 5, 1909. 
Phacoides nuttalliif Conrad, Reagan, p. 179 ; not of Conrad. 

A comparison of the specimens named by Reagan acutilineatus 
and nuttallii shows that they are internal casts of the former 
species. A fragment queried on the label as possibly nuttallii 
appears to be part of the valve of a Protocardia. 

Thyasira bisecta Conrad. 

Thyasira bisecta (Conrad) Reagan, p. 180, pi. 1, fig. 7, 1909. 
This species is correctly identified. 

W. H. Ball — Fossils of the Olympic Peninsula. 307 

Chione vespertina Conrad. 

Cytherea vespertina Conrad, Am. Journal Sci., 1848, p. 432, 

fig. 9. 
Venus (Chione) vespertina Reagan, p. 181, pi. 1, fig. 8, 1909. 

The specimen identified by Reagan as C. vespertina is not in 
the collection. 

Cyclinella 1 sp. 

Venus (Chione) angustifrons (Conrad) Reagan, p. 181, pi. 1, 
fig. 9, 1909. 
This fossil is an internal cast of obliquely oval form, thin 
shelled, and showing a deep sharply angular pallial sinus and 
unusually large posterior adductor scar, suggestive of the genus 
Cyclinella. The figure is extremely inadequate. 

f Venus parapodema Dall. 

Venus (Chione) mathewsoni (G-abb) Reagan, p. 181, pi. 1, fig. 
10, 1909. 
The type specimen is a much dilapidated Venerid shell which 
may possibly belong to V. parapodema, but certainly cannot be 
identified with the Chione. 

Chione securis Shumard. 

Venus (Chione} temblor ensis (Anderson) Reagan, p. 182, pi. 1, 

fig. 11, 1909. 
Venus (Chione) clallamensis Reagan, p. 183, pi. 1, fig. 13, 1909. 
The specimen named V. clallamensis is the internal cast of a 
young specimen of C. securis. 

Antigona olympidea Reagan. 

Venus (Chione) olympidea Reagan, p. 182, pi. 1, fig. 12, 1909. 

The specimen thus named is remarkable for its deep lunular 
depression. The few particles remaining of the external layer of 
the shell indicate a finely reticulate sculpture. The form of the 
cast recalls Venus fordi Yates, of the recent Californian fauna. 

Marcia oregonensis Conrad. 

Saxidomus gibbosus (Gabb) Reagan, p. 184, pi. 2, fig. 14, 1909. 

The identification is correct, but Gabb 's name is a synonym of 
Conrad's Cytherea oregonensis, 1848. 

Macoma calcarea Gmelin. 

Tellina albaria (Conrad) Reagan, p. 184, pi. 2, fig. 15, 1909. 

The type is a specimen of Gmelin 's species which is abundant 
in these deposits, and is quite distinct from albaria. 

308 W. H. Ball — Fossils of the Olympic Peninsula. 

Tellina arctata (Conrad) var. juana Reagan. 

Tellina arctata (Conrad) Reagan, p. 186, pi. 2, figs. 16, 16 a., 

Tellina arctata, var. juana Reagan, p. 186, pi. 2, fig. 17, 1909. 

The specimens are all internal casts of which that designated 
as variety juana is closest to Conrad's type of arctata. It is prob- 
able that the differences all lie within the range of individual 

Tellina oregonensis Conrad. 

Tellina clallamensis Reagan, p. 186, pi. 2, fig. 18, 1909. 

The type specimen is a defective internal cast, apparently 
referable to Conrad's species. 

Spisula albaria Conrad. 

Mactra gibbsana (Meek) Reagan, p. 187, pi. 2, figs. 19 aJ>., 1909. 
This specimen, an internal cast, seems referable to an adoles- 
cent Spisula albaria. 

Panope a brupta Conrad. 

Panopea generosa (Gould) Reagan, p. 188, pi. 3, fig. 20, 1909. 

The specimen exhibits both valves, one imperfect the other 
intact. The draughtsman has selected the imperfect one for 
figuring. The perfect valve agrees exactly with Conrad's type 
of My a (= Panope) abrupt a. 

Teredo sp. indet. 

Teredo sp. Reagan, p. 188, pi. 3, fig. 21, 1909. 

The specimen is a mass of small nearly straight tubes, appar- 
ently of Teredo, with a diameter at the aperture of about a milli- 
meter and a half. 

Xylotrya ? sub striata Conrad. 

Teredo bulbosa Reagan, p. 189, pi. 3, fig. 22, 1909. 

This is identical with Conrad's species. The "bulbs" are the 
proximal ends of adult tubes formed when the borer has com- 
pleted its growth. 

Dentalium I sub striatum Conrad. 

Dentalium snbstriatum (Conrad) Reagan, p. 189, pi. 3, fig. 23, 
The type consists of short pieces from the anterior end of a 
large thin Dentalium, showing only incremental sculpture. The 
specific identification is uncertain. 

W. H. Dall — Fossils of the Olympic Peninsula. 309 

Olivella pedroana Conrad. 

Olivella pedroana (Conrad) Reagan, p. 189, pi. 3, fig. 24, 1909. 
This is a young shell, but seems to be correctly identified. 

% Buccinum clallamensis Reagan. 

Pisiana clallamensis Reagan, p. 190, pi. 3, fig. 25, 1909. 

The specimens are internal casts, probably of a species of Buc- 
cinum. A good species but certainly not a Pisania. 

1 Chrysodomus postplanatus Dall. 

Chrysodomus gettysburgensis Reagan, p. 190, pi. 3, fig. 26, 1909. 
An unidentifiable very defective internal cast, most nearly 
resembling the Astoria species. 

Nucella lima Martyn, var. ! 

Purpura canaliculata (Duclos) Reagan, p. 192, pi. 3, fig. 27, 
The figured specimen is a small individual closely related to 
if not identical with N. lima Martyn. Another specimen not 
figured but similarly labelled is Nucella precursor Dall, described 
from the Astoria horizon at Coos Bay. 

Neverita ? inezana Conrad. 

Polynices {Neverita) recluziana (Petit) Reagan, p. 193, pi. 3, 
fig. 28, 1909. 
These are all decorticated specimens and their identification 
cannot be positive, but they agree fairly well with Conrad's 
species. They are quite distinct from the P. recluziana. The 
Polynices saxea Conrad, mentioned on p. 193, is not among the 
specimens received. 

Amauropsis 1 oregonensis Dall. 

Polynices (Lunatia) olympidea Reagan, p. 194, pi. 3, fig. 29, 
This is also decorticated, it seems close to the Amauropsis, but 
appears to have been a much heavier shell. 

Sinum scopulosum Conrad. 

Sigaretus scopulosus (Conrad) Reagan, p. 194, pi, 3, fig. 30, 1909. 
This is correctlv identified. 

310 W. H. Ball — Fossils of the Olympic Peninsula. 

Trochita inornata Gabb. 

Trochita inornata (Gabb) Reagan, p. 195, pi. 3, fig. 31, 1909. 

This is an imperfect internal cast, but probably belongs to 
Gabb's species. 

Eudolium petrosum Conrad. 

Bolium petrosum (Conrad) Reagan, p. 195, pi. 3, fig. 32, 1909. 
This is correctly identified. 

Eudolium biliratum Conrad. 

Dolium biliratum (Conrad) Reagan, p. 196, pi. 3, fig. 33, 1909. 

This is a badly crushed internal cast, but is probably Conrad's 

Fossils of the Quillayute Fokmation. 

Yoldia (Cnesterium) oregona Shumard. 

Yoldia cooperi (Gabb) Reagan, p. 206, pi. 4, fig. 34, 1909. 
This is not the recent Y. cooperi Gabb, though closely allied. 

Cardium coosense Dall. 

Cardium meekianum (Gabb) Reagan, p. 206, pi. 4, fig. 35, 1909. 
The type specimen comprises an internal cast of both valves of 
what appears to be a young shell of C. coosense; but among the 
miscellaneous duplicate material which came with the types, I 
find a fragment which may be referable to the true C. meekia- 

Cardium corbis Martyn. 

Cardium corbis (Martyn) Reagan, p. 207, pi. 4, fig. 36, 1909. 

The internal cast which served as Reagan's type, appears to 
have been correctly identified. 

Mactra (Spisula) arnoldi n. nom. 

Tivela crassatelloides (Conrad) Reagan, p. 208, pi. 4, figs. 37 a., 

37 o., 37 c, 1909. 
Mactra albaria Arnold, Bull. U. S. Geol. Survey, No. 398, pi. 41, 

fig. 4, 1910 ; not of Conrad. 
The heavy valve referred to Tivela by Reagan, belongs to the 
Mactroid shell figured by Arnold under the name of Albaria, but 
which is much more elevated and triangular than the true albaria. 
Specimens of albaria are rare in Reagan's material, though he 
has identified one as M. falcata Gould. 

W. H. Dall — Fossils of the Olympic Peninsula. 311 

Chione staleyi Gabb. 

Dosinia staleyi Gabb, Pal. Cal., vol. 2, p. 24, 1866. 
Tapes staleyi Gabb, Pal. Cal., vol. 2, part 2, pi. 7, fig. 42, 1868 ; 
Reagan, p. 209, pi. 4, fig. 38 c. only, 1909. 
The specimens figured as number 38 a. and 38 b., are defective 
individuals of Chione securis Shumard. 

Macoma inquinata Deshayes. 

Macoma inquinata (Deshayes) Reagan, p. 209, pi. 4, fig. 39, 1902. 
Compared with specimens of the recent Macoma inquinata var. 
arnheimi Dall, this rather imperfect individual agrees fairly 
well. It is shorter and less acuminate behind than the typical 

Tagelus sp. 

Tagelus calif ornicus (Conrad) Reagan, p. 211, pi. 4, fig. 40, 1909. 
The type is a fragment probably belonging to the genus, but 
insufficient to identify specifically. 

Mactra (Spisula 1) precursor Dall. 

Mactra calif omica (Conrad) Reagan, p. 211, pi. 4, fig. 41, 1909. 
This is the internal cast of a large species whose form is analo- 
gous to that of M. nasuta Gould, though not identical. It is 
numerous in the duplicate material, but has little in common with 
the M. calif omica which is a small species. 

Mulinia olympica n. nom. 

Mactra exoleta (Gray) Reagan, p. 212, pi. 4, figs. 42 a., 1909. 
Mactra hemphilli (Dall) Reagan, p. 212, pi. 4, fig. 43, 1909. 

Though the hinge is not visible, this appears to be a small 
Mulinia, in no way related to M . exoleta. The hinge figured at 
42 b., belongs to a different species. Figure 43 depicts a slightly 
larger specimen. It has no resemblance to the recent M. hem- 
philli or exoleta. 

Mactra (Spisula) albaria Conrad. 

Mactra (Spisula) falcata (Gould) Reagan, p. 213, pi. 5, fig. 44, 
This is the only typical specimen of M. albaria among Reagan's 
types. The specimens doubtfully to M. falcata on p. 214, and 
figure on. plate 5, figures 45 a. to 45 c, are defective immature 
specimens of M. arnoldi. 

312 W. H. Dall — Fossils of the Olympic Peninsula. 

Mya intermedia Dall. 

My a truncata (Lin.) Beagan, p. 214, pi. 5, fig. 46, 1909. 

This is doubtless the same as the recent Alaskan species, though 
the specimens are rather defective, but it is certainly not the 
M. truncata. 

? Cylichna alba Brown. 

Cylichna alba (Brown) Keagan, pi. 5, fig. 47, 1909. 

The type specimen has both ends broken off and the aperture 
invisible, but the remains have the aspect of a Cylichna. 

Antiplanes perversa Gabb. 

Pleurotoma perversa (Gabb) Reagan, p. 215, pi. 5, fig. 48, 1909. 
This appears to be correctly identified, though the specimen is 
slightly stouter than the average recent specimen. 

Lor a sp. 

Bela sanctae-monicae (Arnold) Reagan, p. 217, pi. 5, figs. 49 a-b., 
A fragment of the last whorl of a species of Lora, sculptured 
like Arnold's species, but really unidentifiable. 

Lora miona Dall. 

Belafidicula (Gould) Reagan, p. 217, pi. 5, fig. 50, 1909. 

Analogous to but not identical with L. fidicula, the specimen 
being well preserved, is confidently identified with the recent 

Buccinwn 1 tenebrosum Hancock. 

Buccinum bogachieli Reagan, p. 218, pi. 5, figs. 51 a-b., 1909. 

The portions of the type specimen not concealed by the matrix 
are well preserved, and agree closely with the same portions of 
B. tenebrosum. 

Cymatium (Linatella) pacificum Dall. 

Chrysodomus giganteus Reagan, p. 218, pi. 5, fig. 42, 1909. 
Chrysodomus stantoni Beagan, 1. c, fig. 53. 

The sculpture gives this shell a superficial resemblance to such 
species as Chrysodomus liratus, but there is no doubt of its 
identity with the Coos Bay shell. Figure 53 is based on a decorti- 
cated specimen of the same species. 

W. H. Ball — Fossils of the Olympic Peninsula. 313 

Chrysodomus imperialis Dall. 

Neptunea maxfieldi Reagan, p. 219, pi. 6, fig. 54, 1909. 

This has been well figured both in Professional Paper 59, and 
by Arnold in U. S. Geological Survey Bulletin 398. 

Liomesus f sulculatus Dall. 

Monoceros engonatum (Conrad) Reagan, p. 220, pi. 6, fig. 55, 
The specimens are too imperfect basally to be certain that they 
belong to Liomesus. The Trophon species cited on page 220, is 
not among the specimens received. 

Nucella sp. 

Purpura crispata (Chemnitz) Reagan, p. 121, pi. 6, fig. 57, 1909. 

This is hardly identifiable, owing to its defective condition, but 
is probably a Nucella allied to M. lamellosa Gmellin. 

The specimen identified by Reagan as Purpura lapillus is not 
in the collection, but was certainly not the Atlantic species. 

Nucella ( ? var.) quillayutea Reagan. 

Purpura lapillus var. quillayutea Reagan, p. 221, pi. 6, figs. 
57 a-o., 1909. 
This is closely related to Nucella lamellosa, but is sufficiently 
distinct from any of the recent forms to constitute a good variety 
if not a species. 

Strombiformis washingtoni Reagan. 

Eulima washingtoni Reagan, p. 223, pi. 6, fig. 60, 1909. 
Eulima smithi Reagan, 1. c, fig. 61, 1909. 

This is a good and rather remarkable species of Melanelloid 
type. The aperture not being accessible, the genus is not posi- 
tively determined, but the shell resembles Strombiformis riversi 
Bartsch, from the Pliocene of Santa Monica. 

The difference noted by Reagan between his two species is due 
to the fact that one specimen is decorticated. 

Gyrineum mediocre Dall. 

Ranella marshalli Reagan, p. 222, pi. 6, fig. 62, 1909. 

The type specimen is identical with that from Coos Bay. 

Natica (Cryptonatica) consors Dall. 

Natica (Cryptonatica) clausa (Broderip) Reagan, p. 224, pi. 6, 
fig. 63, 1909. 
This is the smaller common Natica of these beds. 

314 W. H. Ball — Fossils of the Olympic Peninsula. 

Polinices (Lunatia) galianoi Dall. 

Polynices (Lunatia) lewisii (Gould) Reagan, p. 225, pi. 6, fig. 64, 
This is the common larger species of this horizon. It does not 
reach the size of L. lewisii, and has a different contour. 

Purpura foliata Martyn. 

Aporrhais (Arrhoges) quillayutensis Reagan, p. 226, pi. 6, fig. 
This is a cast with some remains of the shell adhering and is 
closely related to if not identical with Purpura foliata Martyn, 
better known as Cerostoma foliatum. 

C. J, Hylander — Mid-Devonian Callixylon. 315 

Aet. XXX. — A Mid-Devonian Callixylon; by C. J. 


In the paleobotanical collections of Yale University 
there have long been some nnstudied sections of a 
Devonian wood from Eighteen Mile Creek, New York. 
These sections were made by Dr. Wieland abont 1900, 
from material collected by 0. C. Marsh in 1860. They 
bear numbers 240, 241, and 242. Since this wood con- 
sists merely in siliceous fragments containing much 
residual carbon, and is otherwise poorly preserved, the 
structure is difficult to make out. Only the transverse 
and radial longitudinal sections are diagnostic. The 
tangential section failed, cutting a region of especially 
poor preservation, with oblique compression. 

Nevertheless, a new species of Zalessky's genus Cal- 
lixylon (7) 1 is indicated, and attention to the plants 
rather than associated vertebrates must bring to light 
better material of the original stem type, supposedly a 
foot or more in diameter. If so, further sectioning must 
disclose the better conserved areas as in various other 
instances of petrified stems. Accordingly, it has been 
considered worth while to give the description which 
follows. The value of these notes rests in the fact also, 
that the fine type from the Indiana Black Shale, Cal- 
lixylon Oweni (6), is at hand for close comparison and 
renders less doubtful features that otherwise would be 
obscure. The camera lucida figures here reproduced are 
accurate to scale, and are not restored, but actual 
drawings of the areas they show. 

Generic Position. 

The grouping of the radial pits in discontinuous masses 
undoubtedly puts the wood of these sections into the 
genus Callixylon. Other characters also are those of 
typical Cordaiteah wood, as described by Penhallow (5), 
or Elkins & Wieland (3). In regard to the specimens 
and their inclusion in Zalessky's genus, Dr. Wieland has 
the following to say : 

1 For Literature references see the end of this paper. 

316 C. J. Hylander — Mid-Devonian Callixylon. 

" Seward in his great textbook (vol. Ill p. 292) says of the 
reference of the Indiana black shale Cordaite to Callixylon — 
'Miss Elkins and Dr. Wieland refer some upper Devonian wood 
from Indiana characterized by a grouping of the circular or 
elliptical bordered pits in the radial walls of the tracheids 
similar to those in Callixylon Trifilievi, which they include 
together with the middle Devonian species of Cordaites New- 
berryi in Zalessky's genus. Though these two American species 
are comparable in the discontinuous arrangement of the tracheal 
pits with the Russian type, the latter is characterized by primary 
xylem strands, a feature not recognized in the American stems ; 
it would seem, then, undesirable to adopt the designation Calli- 
xylon in preference to Dadoxylon unless there is evidence as 
to similar characteristics in the primary xylem. ' 

Is this either a guide to convenient usage or a logical con- 
clusion? It is not well to lay stress on the precise position in 
the middle or upper Devonian for any of these forms, since their 
time range is only inferable. The interesting and decisive point 
is that in middle to later Devonian time there was a cosmopoli- 
tan group of Cordaites with the grouped pits, and such forms 
are known in a typical instance to have the old cryptogamic 
wood. The probability is that all have it. But if any of the 
species referred to Callixylon with reservations well within 
the recognized usages of Paleobotany, were later found to lack 
cryptogamic wood, then a new genus would be indicated, perhaps 
a new family. Moreover, mere reference to Dadoxylon would 
settle nothing, where these recurrently variant forms are con- 
cerned. With or without the cryptogamic wood, there would 
still remain the chance that the leaf or floral characters varied 
strongly from the forms primarily designated as Dadoxylon. 
Like so many of the genera of Paleobotany, Dadoxylon is now 
more a group name than a genus in the purely botanical sense. 
And similarly Callixylon, in the first instance a needed generic 
distinction, must share the same fate of ultimate and convenient 
inclusiveness. Only thus may we avoid the use of over-many 
generic names in our descriptions and groupings of ancient 
plants. The other alternative must be genera of mainly one 
sole species. 

These more or less silicified bits of lignitic wood must have 
attracted the attention of Marsh while searching in the lime- 
stone which thinly covers the Hamilton, and carries both fish 
and plant remains. They would thus be from near the close 
of the mid-Devonian. But the horizon might be in the Hamilton 
shale, or higher in the black shale of the Portage, also carrying 
plants. The bits were marked by Marsh as from the Hamilton 

C. J. Hylander — Mid-Devonian Callixylon. 317 

This Cordaite is not likeh r to be as old as the Paleopitys 
Milleri of the Old Eed of Scotland; 2 and, incidentally, even 
the wood-cuts of the latter given by Hugh Miller in the 'Testi- 
mony of the Rocks,' fig. 3, permit fair judgment of the main 
features. It is seen that the pitting is of the Dadoxylon type 
(2 to 3-seriate), and the rays thin (uniseriate), as Miller 

' The fossil botanist on taking leave of the lower Carboniferous 
beds, quits the dry land and puts out to sea, ' So wrote the stylist 
of Cromarty over seventy years ago ; and while not intended in 
the very severest literal sense, the epigram was long justified. But 
new methods of study and new discoveries have broadened the 
paleobotanic horizon, and helped to give material once thought 
inadequate, high value as evidence of structure, or distribution, 
or both. Perhaps no further answer is needed to any question 
why the fossil tree type here described was not earlier taken up." 


The traclieids of the sections cited are mostly rectangu- 
lar in cross-section, and vary much, in size. The average 
is thirty-five to forty-five microns across, in this respect 
differing sharply from the larger tracheids of Callixylox 
Owexi (3), which run from forty-five to sixty microns 
across. The w r alls appear thick, from three to five 
microns. In radial longitudinal section, the tracheids 
show the chief characteristic of the genus — the bordered 
pits of the radial walls, aligned in discontinuous groups. 
As in Callixylox Owexi (3), the pits are circular or 
irregularly elliptical in outline, and occur in a varying 
number of vertical rows, from one to three (in rare 
instances, four). These vertical rows of pits are rather 
closely set, without compression to marked hexagonal 
form, in the groups as aligned in radial bands that corre- 
spond to the bands of pit groups on the neighboring 
tracheid. Only in a few places in the Eighteen Mile 
Creek material is the grouping of the pits fully visible 
over any great area ; the best area of preservation is 
showm in ^.g. 2. These features can of course only appear 
to advantage where there is little compression and the 
section cuts closely to the true radial wood lines. 

2 McNab, W. E. : On the Structure of a Lignite from the Old Eed Sand- 
stone; Trans. Bot. Soc. of Edinburgh, vol. 10, p. 312. 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 22. — October, 1922. 

318 C. J. Hylander — Mid-Devonian Callixyl 



O t 

o o 

Q c '0 

Fig. 1-6. Explanation opposite. 

C. J. Hylander — Mid-Devonian Callixylon. 319 

In the transverse section, the wood rays appear unusu- 
ally wide, compared with the tracheids. They are also 
of considerable depth, varying from ten to fully twenty 
cells deep. In only one instance, in the radial longi- 
tudinal section, could the end of a ray cell be distin- 
guished, and that is shown in fig. 1. The ray cells appear 
to be about the same size as those of Callixylox Oweni. 
The cell-width of the rays is not visible in the imperfect 
tangential section, but the rays must be in part two-cells 
wide, as in C. Oweni. "Wieland (6, p. 123) notes that in 
Araucarioxylon stems the wood rays are never more than 
two cells wide, although rays three and four cells wide 
are not infrequent in the older Cordaites. 

The growth ring (fig. 5), as also in C. Oweni (fig. 3), is 
a significant feature of this wood, which may have been 
widespread in the mid-Devonian. Such growth rings are 
not present in the Russian Callixylon; nevertheless it is 
likely that the feature is more or less general in Cal- 
lixylon, and it occurs in various other typical American 
Cordaites. Miss Goldring finds it in a Carboniferous 
Cordaite as far south as southern Texas (4). These 
ancient rings are not thought to indicate as sharp a 
seasonal change as the rings in Dicotyledonous plants. 
But on the other hand, little attention has been given to 
the fact that in old and simpler types of wood, growth 
rings do not become a feature of the wood structure. 
The simpler type of growth ring occurs in both Mesozoic 
and recent Cycads, and Chamberlain (1) has now 
observed it in a Monocotyl. Accentuation of growth ring- 
is mainly correlated with the more marked tracheidal and 
ray differention of mid to later Mesozoic time. 

Explanation of Figures 1-6. (All figures enlarged 100.) 
Callixylon Marshii, sp. no v. 

Fig. 1. Eadial long, section showing approximate height of wood ray. 

Fig. 2. Eadial long, section showing the aligned grouping of the radial 

Fig. 5. Transverse section showing growth ring and average appearance 
of tracheids in one of the less compressed areas. 

Fig. 6. Transverse section showing typical wood and conservation. 

Callixylon Oweni, Elkins & Wieland. 

Fig. 3. Transverse section, showing growth ring and larger and well 
preserved tracheids. [Tracheid ends probably aligned.] 

Fig. 4. Eadial long, section, showing radial grouping of pits, for com- 
parison with fig. 2. 

320 C. J. Hylander — Mid-Devonian Callixylon. 

Specific Characters. 

Comparison of the type sections of Callixylon Oweni 
with those of the present Eighteen Mile Creek wood 
sections, reveals specific variation. There are the same 
discontinuous pit groups and obscure growth rings, but 
the tracheids are of a consistently smaller size not 
accounted for by compression or accidental variations, 
and the wood rays are relatively broader. Therefore, 
there need be no doubt in naming it after its finder of 
sixty years ago, who later became so famous as an 
indefatigable collector and paleontologist. 

Callixylon Marshii sp. no v. Tracheids 35-45/* in 
diameter, walls 3-5/* thick, radial pits of the bordered 
crossed slit type, arranged in from one to three vertical 
rows in discontinuous groups of few to 12 or 13 pits to 
the grouping, groupings radially aligned ; wood rays few 
to 20 cells deep, two cells wide, conspicuously broad in 
cross section; growth ring of primitive type present but 
difficult to see. Horizon, upper mid-Devonian of New 
York. Type locality, Eighteen Mile Creek, New York. 

The following comparison is added for convenience : 
Trans. Sect. — Diameter of tracheids : C. Marshii, 35-40/* ; 
C. Newberryi, 44-55/*; C. Oweni, 45-60/*. Thickness of 
walls : C. Marshii, 3-5/* ; C. Newberryi, 6/* ; C. Oweni, 
5/*. Growth rings absent in C. Newberryi, also C. Tri- 
filievi, present in the other two species. Radial long. 
Sect. Pit groups : 3-13 in C. Marshii, 6-13 in C. New- 
berryi, and 3-40 in C. Oweni. Pit diameter: 9.3/* in C. 
Newberryi, 8-10/* in C. Marshii, and 10-11/* in C. Oweni. 
Tang. Sect. Height of rays : C. Marshii, 1-20 ; C. New- 
berryi, medium height; C. Oweni, 1-40. Thickness of 
rays : C. Marshii, at least 2-seriate ; C. Newberryi, 
3-seriate rarely; C. Oweni, 2-seriate. 

Thus it is seen that the C. Marshii is closer to C. New- 
berryi (2) than to C. Oweni. And this is an interesting 
point since the C. Newberryi is given from the mid- 
Devonian of Ohio. Neither of these forms can be con- 
fused with the forms called Badoxylon Ouangondianum 
and D. Halli (2), from the mid-Devonian of New York. 
These are distinct as forms with four and five pit rows, 
the adpressed pits, and broad wood rays. They are men- 
tioned merely because figured by Dawson along with the 

C. J. Hylander — Mid-Devonian Callixylon. 321 

C. Newberryi (2). A form Ormoxylon (2) is given with, 
three pit rows, as having very narrow wood rays one cell 
thick, but this is more likely a typical Dadoxylon. 

Literature cited. 

1 Chamberlain, C. J. : Growth Eings in a Monocotyl, Bot. Gaz., 72, 293- 
304. Text fig. 1-16, 1921. 

2 Dawsox, J. W. : Fossil Plants of the Devonian and Upper Silurian of 
Canada, 92 pp., 20 pis., Geol. Snrv., Canada. Montreal, 1871. 

3 Elkixs, M. G., & Wieland, G. E. : Cordaitean Wood from the Indiana 
Black Shale, this Journal, 38 : 65-78. pi. 1, 2, 1914. 

4 Goldring, W. : Annual Eings of Growth in Carboniferous Wood, Bot. 
Gaz., 72, 326-330, 1921. 

5 Pexhallow, D. P. : Notes on the North American Species of Dadoxylon, 
Trans. Eoy. Soc. Canada, 6, IV, 61, 1900. 

Wielaxd, G. E. : Flora Liasica de la Mixteca Alta, Inst. Geol. Mex., 
Bol. 31, 1914. 

T Zalessky, M. D. : Etude sur 1 'anatomie du Dadoxylon Tchihatcheffi. 
Goppert sp. Mem. du Comite Geol. Nouvelle Ser. Liv. 68, pp. 18-29, Pis. 
I-IV, St. Petersburg, 1911. 

Osborn Botanical Laboratory, 
Yale. University. 

322 Scientific Intelligence. 


I Chemistry and Physics. 

1. A Revision of the Atomic Weight of Beryllium: The Analy- 
sis of Beryllium Chloride. — 0. Honigschmid and L. Birkenbach 
call attention to the circumstance that the atomic weight of this 
metal, usually called glucinum in this country, is of much theo- 
retical interest in connection with the study of the structure of 
atoms. For instance, according to the idea of Rydberg, as it has 
the even atomic number 4, its atom should consist of two helium 
atoms with 8 as its atomic weight, instead of about 9 as now 
accepted, or it might be composed of two isotopes of the values 

8 to 12. Another view is that the atom is composed of two 
heliums and one hydrogen, making 9 the value. The fact 
observed by Strutt that beryl contains appreciable quantities of 
enclosed helium, indicating radioactive transformation, which is 
not known to be the case with any other minerals except those 
containing thorium or uranium, adds interest to the problem of 
this atomic weight. 

The present investigators have made a very careful analysis of 
beryllium chloride, using the most refined modern methods and 
have obtained remarkably closely agreeing, results with an aver- 
age of 9.018. This is 1% lower than 9.1, the value accepted by 
the International Committee. The authors have pointed out a 
very probable source of error in previous determinations where 
the oxide was weighed in the fact that this powder is capable of 
absorbing atmospheric air to an appreciable extent. The fact 
that this atomic weight appears to be so close to the whole number 

9 is of much interest. — Berichte, 55, 4. h. l. w. 

2. Experimental Attempts to Decompose Tungsten at High 
Temperatures. — Gerald L. Wendt and Clarence B. Irion have 
considered the circumstances that atomic disintegration by radio- 
activity has been recognized for 20 years, that this decomposi- 
tion has been artificially produced recently by Rutherford by the 
impact of a-p articles on light atoms, that there is astronomical 
evidence that the heavier atoms do not exist at the temperatures 
of the hottest stars, and that Anderson, two or three years ago, 
devised a method of exploding wires at temperatures above 
20,000°, well above that attributed to the hot stars, a method 
which has become valuable in spectroscopy. They have, therefore, 
carried out the explosion of fine tungsten wires within strong 
glass bulbs, so that the gaseous products of the explosions could 
be collected for analysis. A powerful electric current, usually 
of about 30,000 volts, which passed suddenly across a spark-gap, 
was used to cause the explosions. These were made either in a 
vacuum, or in carbon dioxide. The greatest precautions were 

Chemistry and Physics. 323 

employed to exclude impurities from the bulbs. The principal 
feature of the results is that helium appears to be produced by 
these explosions of tungsten. Several lines in the spectra of 
the resulting gases were not identified, but it appeared that 
hydrogen was not formed. In the 21 recorded explosions in the 
presence of carbon dioxide, the residues after the absorption of 
the latter were remarkably variable, but about 20 cc. of the total 
gas thus obtained was accidentally lost before it was analyzed. 
This report is merely a preliminary one. The senior author has 
been obliged to suspend work for a year or more on account of 
ill-health, and the authors regret the somewhat sensational report 
of their work in the public press last March. — Jour. Amer. Chem. 
Soc. 44, 1887. h. l. w. 

3. A Micro-Method for the Determination of Molecular 
^Yeights in a Melting-Point Apparatus. — Nearly all of the 
organic solvents show only a few degrees of depression in their 
melting points from the presence of one gram-molecule per 
kilogram of dissolved substance, but Karl Hast has found in 
camphor a solvent that is distinguished by such an extraodinarily 
high depression that it presents the possibility of using a ther- 
mometer graduated in whole degrees, instead of the Beckmann 
thermometer, in determining molecular weights, and in permit- 
ting the use of the simple capillary-tube melting-point appara- 
tus for the purpose. The freezing-point depression of camphor 
amounts to 40° for a normal solution (gram-molecules per kilo- 
gram), and since camphor is a solvent which will dissolve many 
organic substances in normal quantity, and almost all in y 2 or % 
normal amounts, the depressions can be accurately measured. 
The method is very simple. A few milligrams of the substance 
are fused with 10 to 20 parts of camphor in a small clean test- 
tube, and after solidification a portion is used for a melting-point 
determination. The original article should be consulted in 
regard to certain details. It appears that this simple method will 
be very useful to organic chemists. — Berichte 55, 1051. 

h. l. w. 

4. ~\Yomen in Chemistry: A Study of Professional Oppor- 
tunities. By The Bureau of Vocational Information. 
Svo, pp. 273*. Xew York City, 1922. (To be obtained from the 
above-mentioned Bureau, 2 West 43d St. Price, postpaid, paper- 
bound $1.10, cloth-bound $1.60.) 

This report has been issued under the auspices of a large and 
distinguished advisory council which includes the presidents of 
the prominent colleges for women in this country. Informa- 
tion has been obtained directly from, and indirectly concerning, 
hundreds of women chemists, while a great many officials of 
scientific, educational and industrial institutions have contrib- 
uted their experiences and opinions in regard to the matter. 

324 Scientific Intelligence. 

The study of the subject has been an exceedingly extensive and 
thorough one, and the results are reported in a very excellent 
manner. There is a very elaborate classification of the different 
positions in chemistry, showing the varied character and wide 
scope of the work. The prejudices to be met by women, the 
training requirements, salaries, means of obtaining positions, and 
many other topics connected with chemical work are well dis- 

The book is to be highly recommended for the perusal of 
women students who are attracted by the study of chemistry and 
are considering making a specialty of the subject. It is an excel- 
lent book also for those who give advice to such students, and it 
may be added that young men who are prospective chemists can 
find in it much valuable information in regard to chemical 
positions. h. l. w. 

5. Spectrum of Aurora. — In making the extended series of 
photographs of all light of the night sky reported in this journal 
3, 476, 1922, Lord Raleigh was fortunate in securing spectra of 
the aurora which show more detail than has been obtained in any 
previous photographs, but, except for the aurora line 5578 of 
unknown origin, the whole spectrum is accounted for by three 
groups of bands in the blue, the violet, and the ultra violet which 
have been proved by Bossekop to be negative nitrogen bands. 

As the origin of this phenomenon has never been fully 
explained Rayleigh 7 s study and attempt to reproduce these bands 
artificially are worth reporting. 

A large tube through which a stream of rarefied nitrogen could 
be passed was prepared and arranged so that the various features 
of the cathode discharge could be isolated and their spectra sep- 
arately photographed. The sources of light were four. (1) The 
pencil of cathode rays, which could be easily deflected by a weak 
magnetic field. (2) The undeflected or so-called "retrograde" 
rays which may be designated as atomic rays in distinction 
from (1) which are electronic. (3) The negative glow. (4) The 
so-called dark space which really emits some light and is only 
dark by comparison. 

It was found that the relative intensities within the blue and 
violet groups of bands vary with the different kinds of excitation 
and that this difference was most conspicuous within the blue 
group. The point under examination was to determine which 
mode of excitation most closely approximated the relative inten- 
sity of these bands in the auroral spectrum. It was found that 
these intensities in the cathode rays, the negative glow and the 
dark space gave no approximation to the auroral conditions but 
the atomic ray spectrum did exhibit about the same intensity 
ratios as the spectrum of the aurora. 

Numerous experiments on variation of pressure were tried as 

Chemistry and Physics. 325 

well as observations on the capillary spectrum-tube of ordinary 
form, at low pressure, but in no case does the development of 
the negative bands by cathode rays imitate the aurora as well as 
the atomic rays. As far as it goes this tends to support the idea 
that the aurora arises from rays from the sun of an atomic nature 
but our author has never been able to secure experimentally 
quite so simple a spectrum as that of the aurora. The negative 
nitrogen bands are accompanied by other details such as bands 
of the positive group or by the line 3995. Traces of hydro- 
gen also are always present in the spectrum, the source of this 
gas evidently being the material of the electrodes. 

It has been presumed on theoretical grounds that the upper 
air in which the aurora occurs is rich in helium and hydrogen but 
neither the present nor previous photographs show any traces of 
these gases in the spectrum of the aurora. The only alternatives 
are either that these gases are absent or that the conditions of 
excitation are not such as to develop their spectrum. As to 
Irydrogen, all laboratory experience shows that any kind of elec- 
tric discharge through air will reveal the smallest trace of hydro- 
gen by its spectrum. The difficulty of explaining the absence of 
helium lines is not so great but it is still serious. On the diffu- 
sion theory the composition of the atmosphere well below the 
measured height of the auroral discharge should be at least five 
volumes of helium to one of nitrogen. If these figures are 
accepted it is very difficult on the hypothesis of atomic ray excita- 
tion to explain the absence of both the helium and the nitrogen 
lines from the spectrum. On the hypothesis of cathode ray 
excitation the helium lines might be vanishingly feeble but a 
new difficulty appears, namely that the intensity distribution in 
the nitrogen band spectrum of the aurora is not that character- 
istic of cathode rays but rather of atomic rays. The author dis- 
misses the matter with the suggestion that the .true mode of 
excitation may be something entirely different from either. — 
Proc. Roy. Soc., 101, 114, 1922. f. e. b. 

6. Suspended Impurity in the Air. — Some of the methods 
hitherto used for the trapping and examination of suspended 
impurities in the air are the following: (a) The filter-paper 
method, in which a given volume of air is filtered through white 
paper leaving a smudge upon it. From the darkening of the 
paper some estimate of the amount of impurity present is 
afforded by comparison with a properly constructed scale of 
shades. It is applicable only to the case of dark colored particles 
and their quantitative estimation is difficult as the amounts col- 
lected are hardly weighable. It has the further drawback that 
the sediment thus entangled in the meshes of the paper cannot be 
subjected to microscopic examination. 

(b) The condensation method. When air not too dry and con- 

326 Scientific Intelligence. 

taining foreign particles is suddenly expanded condensation 
occurs about these particles which subsequently settle down in 
drops of moisture upon a glass floor where they may be counted. 
As water is known to condense on ions as well as upon dust par- 
ticles the results obtained by this method cannot be very exact. 

(c) The soluble filter method. When soluble filters such as 
collodion-wool or sugar are used the filter may be dissolved and 
the number of trapped particles in a small measured volume of 
the solution counted. The difficulty with this method is that the 
suspended particles may themselves be soluble, or particles which 
were aggregates would tend to subdivide, (c) The spray 
method. The air may be drawn through a spray and the cap- 
ture counted or weighed after evaporating the water. As it is 
extremely difficult to wet dry dust this method cannot be con- 
sidered very efficient. 

It is to be noted that while all these methods tell something as 
to the matter deposited the relation as to the amount actually in 
suspension is rather uncertain. An improved method devised by 
Dr. J. S. Owens is free from many of the above objections. He 
has found that when a very small jet of air properly humidified 
is made to impinge at a high velocity upon a surface such as a 
microscope cover glass, the foreign particles adhere to the glass 
and a record is obtained which may be removed, examined, and 
counted microscopically. By varying the conditions of the 
experiment he was able to select the form of the orifice, the 
velocity of the air and arrangement of apparatus which gave the 
most satisfactory results. Also by subjecting a specimen of air 
to dust extraction in successive cells a very definite knowledge of 
the efficiency of the apparatus wias secured. For the details the 
reader will naturally refer to the original paper which also con- 
tains an extended record of observations made at Norfolk, Eng- 
land, and a discussion of the probable sources of contamination. 
This investigation was carried out for the Advisory Committee 
on Atmospheric Pollution. — Proc. Boy. Soc, 101, 18, 1922. 

F. E. B. 

7. The Principles of Geometry ; by H. F. Baker, vol. I, pp. 
VIII, 183. Cambridge, 1922 (Cambridge University Press).— 
The purpose of the author is to place the reader in touch with the 
main ideas dominant in contemporary geometry, or at least for 
those parts which precede the theory of higher plane curves and 
irrational surfaces. The present volume is entitled Foundations 
and is devoted to the indispensable logical preliminaries. Its 
method, however, is very different from the usual text book of 
geometry, on the one hand, and the work of the logical rigorist, 
on the other, in which the fundamental conceptions are analysed 
into a sufficient and final number of axioms. Adopting the view 
that a Science grows up from the desire to bring the results of 

Miscellaneous Scientific Intelligence. 327 

observations or the relations of a class of facts which appear to 
be connected under as few general propositions as possible, if it 
is found necessary to introduce abstract entities transcending* 
the actual observations, the author is of the opinion that the 
degree of abstractness which may be usefully and safely applied 
is a matter for judgment and choice. 

In particular the consideration of distance and congruence are 
rejected as fundamental ideas but in effect they are replaced by 
a theory of related ranges. Although the plan soon leads into 
the exposition of non-Euclidean geometry, the author believes 
that many students of the class who look forward to becoming 
engineers or physicists will find the course stimulating and easy. 

Subsequent volumes will deal on the basis of results obtained 
in this volume, with conies, quadric surfaces, cubic curves in 
space, cubic surfaces and certain quartic surfaces. f. e. b. 

IT. Miscellaneous Scientific Intelligence. 

1. Smell, Taste, and allied Senses in the Vertebrates; by 
G. H. Parker. Pp. 192, with 37 figures. Philadelphia and 
London, 1922 (J. B. Lippincott Company). — This new volume in 
the series of monographs on experimental biology by American 
writers explains without unnecessary technicalities the structure, 
physiology, embryological development, and evolution of the 
chemical sense organs, or chemoreceptors, of man and the other 
vertebrate animals. The subject is treated from a comparative 
standpoint, the conditions found in man being first discussed, 
followed by an account of the homologous structures in the other 
groups of animals. The final chapter explains the interrelations 
of all the chemical senses and compares the chemoreceptors with 
the receptors for mechanical and radiant energy. At the end 
of each chapter is a brief bibliography of the more important lit- 
erature on the subject. 

The book brings together the results of recent studies on these 
complex sense organs, some of the most important of which have 
come from the author's own investigations, and makes the infor- 
mation so comprehensible that even the general reader can now 
gain a clear conception of the nature of some of the sensations 
which he has daily experienced but little understood, w. r, c. 

2. Science and Human Affairs from the Viewpoint of Biol- 
ogy ; by \Yintertox C. Curtis. Pp. vii, 330. New York, 1922 
(Harcourt, Brace and Company). — The author follows the his- 
tory of the scientific method from the earliest times and shows 
how human progress has always been associated with advances in 
natural science. And this is particularly true of biological 
science, by means of which man has come to know about himself, 
his origin, his developmental history, his physical, physiological 

328 Scientific Intelligence. 

and psychical peculiarities, with the gradual emancipation of 
mankind from persecution and superstition. 

The second part of the book treats of the cell doctrine, organic 
evolution, and current problems and methods of zoological 
science, showing the profound influence which recent biological 
discoveries are having upon the philosophy of the everyday life 
of all classes of the people. 

In the third part into which the subject is conveniently divided 
the importance of science in the solution of our present social 
problems is discussed and the dependence of the material and 
spiritual welfare of the' race on continued scientific research is 
emphasized. The author holds that the cure for the present 
widespread social unrest and the solution of the complex prob- 
lems in this reconstructive period will rest upon the recognition 
of nature's laws as the guiding principle in the ordering of all 
human affairs. 

The book is thoughtful and scholarly, well arranged and pleas- 
ingly written ; its reading will give one a more dignified view of 
science as applied to human welfare, with suggestions as to how 
he may aid in social progress. w. r, c. 

3. The Biology of the Sea-shore; by F. W. Flattely and 
C. L. Walton, with an introduction by J. Arthur Thomson. Pp. 
xvi, 336, with 16 plates and 23 text-figures. New York, 1922 
(The Macmillan Company) . — The first two decades of the present 
century may be looked upon as the period of experimental biol- 
ogy, when the most important discoveries have been made in the 
laboratory. More and more, however, in recent years there has 
been a return to out-of-doors nature and a reawakened interest 
in old-fashioned natural history, now refined and dignified by the 
term "ecologj^. " 

This book is one of the first in several years to deal with the 
associations of organisms found along the shores of the seas, but 
the subject is treated in a much broader way than would have 
been possible without the knowledge gained by the intensive 
laboratory studies of past years. 

How the hundreds of species of animals and plants to be found 
in even a limited area between tide-marks are associated together, 
how their distribution is determined by the physical characteris- 
tics of the shore, how the animals are provided with means for 
attack and defense, how they may regenerate their missing parts 
after injury or spontaneous fragmentation, how some forms are 
dependent upon others for food and protection, the wonderful 
adaptations for locomotion of the various types for securing food, 
for respiration under different conditions, their nervous organ- 
ization and behavior, their reproduction and growth, their influ- 
ence on the human welfare — these are the interesting topics 

Obituary. 329 

discussed in so entertaining a manner that one is tempted to 
hurry to the shore and verify the stated facts by first-hand 
observation. w. r. c. 

4. New Meteoric Iron from Kentucky; G. P. Merrill (Com- 
municated). — The National Museum has received from Professor 
A. M. Miller of the University of Kentucky an heretofore unre- 
corded 15 lb. mass of meteoric iron found while plowing in a 
field near Glascow in that state. The iron (an octahedrite) is 
much oxidized and evidently has lain long in the soil. There is 
no record of its fall. This is the eleventh iron meteorite thus 
far found within the state limits. 


Dr. Kollin D. Salisbury, professor of geography and geol- 
ogy in the University of Chicago and dean of the Ogden Grad- 
uate School of Science, died on August 15 at the age of sixty-four 
years. He was graduated at Beloit College in 1881 receiving the 
degree of Ph.B., and later those of A.M. and LL.D. He was also 
professor of geology at the same institution, from 1884 to 1891. 
In 1891-92 he was professor in the University of Wisconsin, since 
which time he has been connected with the University of Chicago.- 
His chief interest was early in glacial geology and subsequently 
more particularly in geography. In glaciology he was connected 
with the Geological Survey of New Jersey and the United States 
Geological Survey. A work in which he took an important part, 
in association with Professor T. C. Chamberlin, was the Geology 
published in three volumes (1904 and 1906). The first volume 
(654 pp.) dealt with Geologic Processes and their results; the 
second and third (692 and 624 pp.) with Earth History. 

In his special subjects Professor Salisbury stood in the front 
as a thinker, investigator and teacher. His death robs the Uni- 
versity of Chicago and the science of the country of a man of 
rare gifts and attainments, and an especially effective teacher. 

Dr. G. H. Cox, recently geologist for the Josey Oil Company, 
and for a number of years professor of geology at the Missouri 
School of Mines, was killed in an automobile accident near 
Bristow, Oklahoma, on August 20. 

Ward's Natural Science Establishment 

A Supply-House for Scientific Material. 

Founded 1862. Incorporated 1890. 

A few of our recent circulars in the various 
departments : 

Geology : J-32. Descriptive Catalogue of a Petrographic Col- 
lection of American Rocks. J- 188 and supplement. 
Price-List of Rocks. 

Mineralogy : J-220. Collections. J-238. Minerals by Weight. 
J-224. Autumnal Announcements. 

Paleontology: J-201. Evolution of the Horse. J-199. Palae- 
ozoic index fossils. J-115. Collections of Fossils. 

Entomology : J-33. Supplies. J-229. Life Histories. J-230. 
Live Pupae. 

Zoology: J-223. Material for dissection. J-207. Dissections 
of Typical Animals, etc. J-38. Models. 

Microscope Slides: J-189. Slides of Parasites. J-29. Cata- 
logue of Slides. 

Taxidermy: J-22. North American Birdskins. Z-31. General 

Human Anatomy: J-37. Skeletons & Models. 

General : J-228. List of Circulars & Catalogues. 

Ward's Natural Science Establishment 

84-102 College Ave., Rochester, N. Y., IT. S. A. 


number consisting of 100 to 120 pages). Editor : EUGENIO RIGNANO. 

This is the only review which has a really international collaboration ; which is 
of world-wide circulation and occupies itself with the synthesis and unification of 
knowledge, in the history of the sciences, mathematics, astronomy, geology, physics, 
chemistry, biology, psychology and sociology. 

It is the only review which, by inquiries among eminent scientists and writers 
(on: The philosophical principles of the sciences; Fundamental astronomical and physical 
questions of current interest; Contributions given by the various countries to different 
branches of knowledge; Question of vitalism; Social question; International questions 
raised by the world war), makes a study of the problems interesting scientific and intel- 
lectual circles throughout the world. 

It has published articles by Messrs. : Abbot- Arrhenius =Ashley = Bayliss - Beichman- 
Bigourdan - Bohlin - Bohn = Bonnesen - Borel - Booty - Bragg = Bruni = Burdick = Carver - 
Caullery - Chamberlin - Charlier = Claparede - Clark - Costantin - Crommelin = Crowter = 
Darwin ■» Delage = De Vries - Durkheim = Eddington = Edgeworth - Emery = Enriques - 
Fabry - Findlay - Fisher = Fowler - Golgi = Gregory - Harper - Hartog - Heiberg - (links - 
Hopkins-lnigues-Innes=Janet-Kaptein-Kaye=Kidd-Langevln-Lebedew=Lloyd Morgan - 
Lodge - Loisy = Lorentz - Loria -Lowell - MacBride - Meillet -Moret-Mulr=Peano -Picard - 
Polncare - Pui seux = Rabaud = Rey Pastor = Righ i = Rignano=Russell=Ru t her f ord=Sagnac - 
Sarton- Schiaparelli - Scott = See - Sherrington - Soddy - Starling- Svedberg- Thomson - 
Thorndike-Turner -Volterra -Webb-Weiss -Zeeman and more than a hundred others. 

"SCIENTIA" publishes its articles in the language of its authors, and joins to the 
principal text a supplement containing the French translations of all the articles that 
are not in French, f Write for a Specimen Number to the General Secretary of 
"Scientia", Milan, sending, -to defray postal and other expenses, -2 Francs in 
stamps of your country.) Annual subscription : 40 sh., or 10 dollars post free. 

Office : 43 Foro Bonaparte, Milan, Italy 

Publishers: WILLIAMS & NORG ATE - London ; FELIX ALCAN - Paris 


Obituary. -K. D. Salisbury: G. H. Cox, 329. 

Art. XXVI. — Jones's Criticism of Chamberlin's Ground- 
work for the Study of Megadiastrophism ; by T. C. 
Chamberlin, 253 

Art. XXVII. — Relation of Sna Water to Ground Water 

along Coasts ; by J. S. Brown, 274 

Art. XXVIII.— A Petrologic Study of the Cape Neddick 

Gabbro ; by A. Wandke, 295 

Art. XXIX. — Fossils of the Olympic Peninsula; by W. II. 

D all, 305 

Art. XXX.— A Mid-Devonian Callixylon; by C. J. Hy- 

LANDER, 315 


Chemistry and Physics. — A Eevision of the Atomic Weight of Beryllium: The 
Analysis of Beryllium Chloride, 0. Honigschmid and L. Birlenbach: 
Experimental Attempts to Decompose Tungsten at High Temperatures, 
G. L. Wendt and C. B. Iron, 322. — A Micro- Method for the Determina- 
tion of Molecular Weights in a Melting-Point Apparatus, K. East: 
Women in Chemistry: A Study of Professional Opportunities, The Bureau 
of Vocational Information, 323. — Spectrum of Aurora, 324. — Suspended i 
Impurity in the Air, 325. — The Principles of Geometry, H. F. Baker, 326. 

Miscellaneous Scientific Intelligence. — Smell, Taste, and allied Senses in the 
Vertebrates, G. H. Parker: Science and Human Affairs from the View- 
point of Biology, W. C. Curtis, 327. — The Biology of the Sea-Shore, 
F. W. Flattely and C. L. Walton, 328.— New Meteoric Iron from Ken- 
tucky, G. P. Merrill, 329. 

Library, U. S. Nat. Museum. 



Established by BENJAMIN SILLIMAN in 1818. 




Editoe: EDWARD S. DANA. 



of Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 


of Washington. 


^ — ^7r*^ 

No. 23 -NOVEMBER, 1922. 


19 22. 


Published monthly. Six dollars per year, in advance. $6.40 to countries in the 
Postal Union ; $6.25 to Canada. Single numbers 50 cents; No. 271, one dollar. 

Entered as second-class matter at the Post Office at New Haven, Conn., under the Act 
of March 3, 1879. 


The Works of 

Q. Montague Butler, e.m., 

Dean, College of Mines and Engineering, 
University of Arizona. 

Geometrical Crystallography 

Gives the reader the ability to recognize crystal forms and especially 
systems at sight, and with the use of few, if any, instruments. 
155 pages. 4 by 6f . 107 figures. Cloth, $1.50. 

A Pocket Handbook of Minerals. 

Gives all the details needed to identify most of the minerals which mining 
men, students, or collectors are apt to encounter. The emphasis is always 
placed upon characteristic physical features. 

311 pages. 4 by 6f. 89 figures. Flexible binding, $3.00. 

Pocket Handbook of Blowpipe Analysis. 

Complete directions are presented for the determination of the composition 
of minerals. 

80 pages. 4* by 6f. Cloth, $1.25. 

Crystallography, Minerals and Analysis, Complete, 

A handy combination of the three above books in one volume. 
546 pages. 4 by 6f. 196 figures. Flexible binding, $4.00. 

Hayes' Handbook for Field Geologists. 

Third Edition, Thoroughly Revised and Rearranged, by SIDNEY 


A well-known educator says, of this book: ''This work has 
been used by me and my students in geological surveying since 
its first appearance .... Dr. Paige has happily preserved with 
improvements the quality and the design of the original work." 

166 pages. 4$ by 6f. 20 figures. Flexible binding, $2.50. 


432 Fourth Avenue 
New York 

London Montreal, Quebec 

CHAPMAN & HALL, Ltd. Renouf Publishing Company 

AJS 11.22 




Art. XXXI. — The Silicates of Strontium and Barium; 
by Pextti Eskola. 

Methods of working 

The system SrO-Si0 2 

The system BaO-SiO. 

The system CaO.SiCX-SrO.SiO, 

The system CaO.Si0 2 -BaO.Si(5 2 

The properties of the glasses. 

Absence of diopside analogs. 

The strontium and barium feldspars. 
Some general considerations regarding the relations of the alkaline earth 



Among the compounds of the alkaline earth metals 
(calcium, strontium, and barium), the sulphates and 
carbonates are common as minerals. In the case of the 
silicates, however, we find those of calcium common, but 
natural silicates of strontium and barium are few in 
number, complicated in composition, and of rare occur- 
rence. Considering the existence of isomorphous carbo- 
nates, sulphates, and other compounds of these three ele- 
ments, the non-occurrence of analogous silicates seems 
striking and suggests experimental investigation to 
answer the questions : Do strontium and barium under 
experimental conditions form such compounds as are 
known in the case of calcium? And, if they do, what are 
their properties and relations to the calcium silicates? 

The calcium silicates which form from dry melts have 
been studied, but very little is known about the strontium 
and barium silicates. G. Stein 1 prepared some silicates 
and determined their melting points as follows : SrSi0 3 , 
1287°; BaSi0 3 , 1368.5°; Sr 2 Si0 4 , 1593°. R. Wallace 2 

1 G. Stein, Z. anorg. Chem., 55, 159, 1907. 

2 B. Wallace, Z. anorg. Chem., 63, 1, 1909. 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 23. — November, 1922. 

332 P. Eskola — Silicates of Strontium and Barium. 

carried out thermal investigations by the cooling curve 
method on the binary systems of the metasilicates of 
sodium and lithium with those of strontium and barium. 
He gives the melting point of SrSi0 3 , 1529° and of 
BaSi0 3 , 1490°. All his melting diagrams resulted in the 
type of complete solid solutions with a minimum. 
P. Lebedew, 3 for the system BaSi0 3 -CaSi0 3 , also found a 
melting curve of the complete solid solution type with the 
melting point of BaSi0 3 at 1438° and a minimum at about 
1000°. Smolensky, 4 who studied the system BaSiO.- 
BaTi0 3 , found the melting point of BaSi0 3 at 1470°. 

All these results were obtained by the cooling-curve 
method, using carbon crucibles and porcelain tubes to 
protect the thermo-elements. No quantitative optical or 
other physical diagnostics for the crystalline phases were 
given. In the absence of such characteristics, and in 
view of the poor agreement of the results, the melting 
diagrams merely based upon cooling curves are subject 
to large corrections and often entirely erroneous. The 
melting point determinations are, in fact, all considerably 
too low. 

Very accurate determinations were carried out by 
Jaeger and Van Klooster, 5 who found the melting point of 
SrSi0 3 , 1578 ± 1°, and of BaSi0 3 , 1604 ± 0.5°. They, 
also, emphasize the unreliability of the cooling-curve 
method and illustrate this by experimental evidence. 
They determined further some physical constants of the 
compounds named. Their results will be mentioned 

The entire field of the physical chemistry of the 
strontium and barium silicates being open, the first task 
was to investigate the two binary systems SrO-Si0 2 
and BaO-Si0 2 . Among the further problems, those 
regarding the relations of the metasilicates to the well- 
known wollastonite minerals seemed to me most interest- 
ing and were chosen for study. For still further com- 
parison with corresponding calcium compounds I tried 
to synthesize the strontium and barium compounds analo- 
gous to two important lime silicate minerals, diopside 
and anorthite. 

3 P. Lebedew, Z. anorg. Chem., 70, 301, 1911. 
4 S. Smolensky, Z. anorg. Chem., 73, 293, 1912. 

5 F. M. Jaeger and H. S. Van Klooster, Proc. Kon. Akad. Amsterdam, 6, 
XVIII, 896, 1915. 

P. Eskola — Silicates of Strontium and Barium. 333 

Methods of Working. 

The mixtures were made up of pure quartz, specially 
prepared calcium carbonate, barium carbonate " Squibb \s 
reagent" and strontium carbonate from "Baker's 
analyzed chemicals." These substances were dried at 
150 : C before weighing, and mixed together, melted if 
possible, chilled and crushed, then reheated and ground 
twice more. 

The equilibrium relations were ascertained almost 
exclusively by the quenching method. Heating curves 
were run in some cases for the purpose of checking the 
temperature measurements. 

The temperatures were determined by potentiometer 
and a Pt-PtRh thermoelement calibrated against the 
melting points of diopside and anorthite. The tempera- 
tures, for the most part, are not far from the diopside- 
point, so that the calibration against diopside, together 
with the temperature limits taken in the quenchings, show 
the actual degree of accuracy of the work. Therefore 
those calibrations are stated below. 

Diopside, melting point 1391.5°, in the standard scale 
corresponds to 14230 microvolts. 

Date (1921) Crystals Melt Correction The systems under 

ix v fi v investigation 

6 April 11100 11150 -4- 105 ") q n ~ n 

28 April 11110 11130 + 105 ) foru-biU 2 

11 Mar 14120 11130 -f 105 BaO-SiO, and SrSiO s -CaSi0 3 

17 June 11090 11110 -f 130 ] 

26 Junea 11310 11330 — 90 

30 June 11310 11330 — 90 ]■ BaSiO-CaSiO, 

5 August 11290 11310 — 70 1 

22 August 11270 14290 . — 50 J 

a Xew thermoelement. 

Iii the determination of the refractive indices I had the 
advantage of using the improved immersion method as 
worked out by Merwin. 6 This method involves an 
improvement in accuracy as well as in completeness, 
making it possible to determine at the same time disper- 
sion as well as refractive indices. One determines 
directly, using a monochromatic illuminator, the wave- 
lengths for which the refractive indices to be measured 

8 E. Posnjak and H. E. Merwin, The ternary system Fe.O-SO-H.O, J. Am. 
Chem. Soc, 44, 1965, 1922. 

334 P. Eskola — Silicates of Strontium and Barium. 

match those of two or more members in the set of refrac- 
tive liquids. The dispersions of the whole series of 
liquids used having been determined and expressed 
graphically, it now remains simply to locate, on the dia- 
gram, the points determined and to read the refractive 
index for any wave-length desired. The values obtained 
in favorable substances are dependable in the third deci- 
mal place, provided due care is taken for variations in the 

In the present work the indices of refraction are given 
for. four wave-lengths corresponding with the spectral 
line F(a = 486/*/*), T1(A = 535^),- D( = 589^) and 
C(A = 656/*/*). Jn the actual determinations wave- 
lengths between 500 and 625 were commonly used, there- 
fore the values for F and C may sometimes have an 
uncertainty of ± 0.001 or a little more, due to enlarge- 
ment of the errors in extrapolation. 

The System SrO-Si0 2 . 

The results of a thermal study of this system are given 
in table I. 

Table I. 



wt. per 


Temperature Time Eesulting phases 


Si0 2 

°C. minutes 

Liquidus of eristobalite 



1636 15 Glass 



1600 60 Glass and eristobalite 
■Eutectic tridymite-SrSi0 3 



1368 ... 30 Glass 



1361 60 - Glass and a little tridymite 



1356 60 Tridymite, SrO.Si0 2 , trace of 
. glass 



1363 25 SrO.Si0 2 and glass 
Liquidus. of SrO.SiO, 



1430 30 Glass 



1420 60 Glass and SrO.SiOo 



1571 60 Glass 



1552 45 Glass and SrO.SiO, 
Melting point of SrO.Si0 2 



1584 15 Glass 



1580 ' 15 SrO.Si0 2 and glass 



1575 15 SrO.SiO, only 

P. Eskola — Silicates of Strontium and Barium. 335 

















o- — 


h — 

2SrO.Si6 2 

Eutectic SrO.Si0 2 -2SrO.Si0 2 * 
1554 15 Glass 

1546 30 Glass and SrO.SiO. 

1538 15 SrO.Si0 2 and 2SrO.SiO, 

1544 15 SrO.Si0 2 and 2SrO.SiOo 

Liquidus of 2SrO.Si0 2 
1617 15 Glass 

1600 15 Glass and 2SrO.Si0 2 

1628 15 Glass and 2SrO.SiO, 

1634 5 Sr 2 SiO, only 

The equilibrium diagram (fig. 1) is based on the above 
results ; the melting point of cristobalite has been placed 
at 1710°, according to the determinations of Ferguson 
and Merwin, 7 and the inversion point cristobalite-tridy- 
mite at 1470° according to Fenner. 8 . 

The liquidus curve of the silica minerals has the same 
general shape as formerly found in all the other binary 
systems with silica as one component : starting from the 
eutectic point first a very steep rise a little above 1600°. 

The eutectic tridymite-SrO.Si0 2 . — I prepared inciden- 
tally a mixture of 46.18 per cent SrO and 53.82 per cent 
Si0 2 , corresponding to the proportion Sr0.2Si0 2 , to find 
out whether there was any disilicate of strontium, analo- 
gous to the barium disilicate known formerly. The 
result was negative, and this happened to be almost 
exactly the eutectic composition. The tridymite liquidus 
lying only about five degrees above the solidus, the 
eutectic composition can be located with a fair degree of 
accuracy at 46.5 per cent SrO and 53.5 per cent Si0 2 . 
The temperature is 1358 ± 4°. 

Strontium metasiiicate, SrO.Si0 2 . — Strontium metasili- 
cate was found in only one form which shows a very close 
resemblance to a-CaSi0 3 , or pseudowollastonite. Its 
melting point was determined as 1580 ± 4°, in agreement 
with Jaeger and Van Klooster's result, 1578 =1= 1°. It is 
apparently of dihexagonal pyramidal symmetry, uniaxial 
and positive, and its characteristic habit is thick-tabular 
.parallel to (0001). There is a fairly good basal cleavage. 

7 J. B. Ferguson and H. E. j&erwin, this Journal, 46, 417, 1918. 

8 C. N. Fenner, this Journal, 36, 337, 1913. 

336 P. Eskola— Silicates of Strontium and Barium. 

Fig. 1. 



500 - 

1400 - 


1 1 

\ i 

1 1 1 1 1 1 




-G \ 












K c 





\ ° 

















\ / Cristobalite 




\ 1 and liquid 



TF \ 




cr \ 



— \ 



\ / 


"^ \ / 




5 \ 




* \ / 





io -A 
o \ 

Tr i d y m it c 
Dud liquid 



i- \ 


in \ 







■*■ V 





5r0-5\0n and tri dymitc 


i i 


I i l 1 1 i 


SL5r0-5iO9 5rO5i0n 

Fig. 1. — Temperature-concentration diagram of the binary system 
Sr0-Si0 2 . 

The density, at 30°, determined with a pycnometer, is 
3.650. This is calculated as real density, compared with 
water at 4° and making correction for the buoyancy of 
air. Jaeger and Van KLooster 9 found d 4 =3.652 at 
25.1°, in good agreement with that found by me which 
gives d 4 =3.653. 

9 Log. cit. 

P. Eskola — Silicates of Strontium and Barium. 337 

The refractive indices were determined in granular 
crystals obtained by cooling a pure melt. They were 
found to be : 

a y 

F 1.606 1.646 

Tl 1.602 1.641 

D 1.599 1.637 

C 1.596 1.634 

These values were repeatedly checked and are correct in 
the third decimal place. Jaeger and Van Klooster give 
less exact values : n x = 1.590 ± 0.003, and n, = 1.620 ± 
a-CaO.SiO 2 has 10 a(D) = 1.610; /3(D) = 1.611; 7 (D) =1.654. 

In the apparent uniaxial positive character and hex- 
agonal tabular habit the strontium metasilicate is similar 
to a-CaO.Si0 2 and, as stated later, they form a continuous 
series of solid solutions. It would seem necessary, there- 
fore, that they should be closely isomorphous. Now 
Wright 11 has proved a-CaO.Si0 2 to be really monoclinic, 
showing, in plates perpendicular to the optic normal, a 
polysynthetic twinning along the basal plane. The optic 
axial angle, though small, is not quite 0°. 

Much care was taken to find out whether the strontium 
metasilicate shows a similar deviation from hexagonal 
symmetry, but no such phenomena were noticed, although 
apparent twinning was observed in the mix crystals of 
SrO.SiOo and CaO.SiO, (cf. below). 

Euhedral crystals of SrO.Si0 2 were obtained from a 
strontium chloride melt with an excess of silica (cf. 
below). They are all thick tabular and apparently 
dihexagonal pyramidal (hemimorphic), one end being 
terminated only by the basal plane (fig. 2), sometimes 
with additional narrow pyramid faces, while the other 
end regularly shows well developed pyramids, usually 
combinations of two, three, or four simple forms. The 
pyramid faces are striated parallel to the edge with the 
basal plane. The greater part of the crystals are 
twinned on the base (fig. 3). 

As the crystals are perfectly hexagonal in aspect and 

10 G. A. Kankin and F. E. Wright: The ternary system CaO-ALO-SiO,, 
this Journal, 39, 1, 1915. 

11 E. T. Allen, W. P. White, and F. E. Wright: On wollastonite and pseu- 
dowollastonite, — polymorphic forms of calcium metasilicate, this Journal, 21, 
89, 1906. 

338 P. Eskola — Silicates of Strontium and Barium. 

the measurements failed to reveal any deviation from 
hexagonal symmetry, I shall provisionally describe them 
as hexagonal, although the isomorphism with the calcium 
metasilicate makes it probable that the strontium metasi- 
licate is pseudohexagonal and really monoclinic, in that 
case belonging to the domatic, or monoclinic hemihedral 

Fig. 2. 

Fig. 3. 

Fig. 4. 

Figs. 2, 3, 4.— Crystals of SrO.Si0 2 . 

In examining these crystals with a binocular micro- 
scope one was found that was larger (about 0.5 mm in 
diameter) and better than the rest, but it was incomplete, 
having only faces in two pyramidal zones developed. 
These zones were identical, each combined of three 
different pyramids, p( 1011), o(4045) and w(2021) (fig. 4). 
While searching for evidence of monoclinic symmetry I 
measured a number of crystals, but unfortunately they 
were too imperfect to prove anything positively. Often 
the pyramidal faces were quite curved in their zones so 
that the signals appeared as lines instead of points. All 
the forms observed in the best crystals were, however, 
also identified in other crystals, and some of them at the 
same_ time in three pyramidal zones. A further form 
t (5051) was found to be very common. 

The following data are based on the best crystal, except 
for the last named form t. 

The crystals separating from silicate melts and 
embedded in glass usually were bipyramidal in habit, 
with large basal planes, and thinner than those obtained 
from the chloride melt. 

In summarizing the evidence regarding the crystal 
form of SrO.Si0 2 , we may state that, from its solid 
solubility and apparent isomorphism with a-CaO.Si0 2 , it 
might be expected to be monoclinic, but actually the 
crystals agree so closely with the dihexagonal pyramidal 

P. Eskola — Silicates of Strontium and Barium. 339 







Limits Average 

0° 0' 
49° 32' 
55° 41' 
66° 55' 

80° 20' 

30° 10' 

c (0001) 

p(10ll) 49° 10' - 49° 55' 49° 32' 
o (4945) 55° 50' — 56° 2' 55° 56' 
n (2021) 66° 32° - 66° 55' 66° 44' 
t (5051) 78° 20' - 80° 50' 79° 17' 

30° - 

Axial ratio a : c = 0. 

symmetry that, judging only from the characters of the 
rather poor crystals available, it should belong to this 

Another question that arose from the similarity to 
a-CaO.Si0 2 was whether SrO.Si0 2 would also appear in 
more than one form. Several experiments were made to 
settle this question. 

SrO.Si0 2 -glass was held one hour at 1020°. The result 
was a crystallized mass showing radiating scaly crystals, 
uniaxial, positive, o = 1.600 ± 0.003, e == 1.640 ± 0.003. 
It is consequently identical with the crystals separating 
from the melt. 

The same result was achieved when the glass was held 
4 hours at 950°. 

SrO.Si0 2 was melted together with SrCl 2 in a Fletcher 
furnace and allowed to cool very slowly. After dissolv- 
ing the chloride in water to which finally was added a 
little HC1 to decompose the chloro-silicate that had 
formed, the result was the same form of SrO.Si0 2 as 

The last-named experiment was modified so that the 
SrO.Si0 2 -SrCl 2 -melt was left in the platinum resistance 
furnace over night at about 1000°. The product was not 
metasilicate at all, but strontium orthosilicate, which was 
identified by its refringence and characteristic twinning. 

Now, imitating the method that Allen and White 12 
found most effective in changing pseudowollastonite into 
wollastonite, I prepared strontium vanadate, Sr(V0 8 ) 2 , 
and melted one gram of this with two grams of SrO.Si0 2 . 
The mixture was held over night at about 900°. After 
washing away the vanadate, the mass consisted of the 

12 E. T. Allen and W. P. White, this Journal, 21, 89, 1906. 

34:0 P. Eskola — Silicates of Strontium and Barium. 

same form of SrO.Si0 2 , as rounded crystal grains and 
with no euhedral forms. 

To prevent the formation of orthosilicate in the stron- 
tium chloride melt I finally heated SrO.Si0 2 with SrCl 2 
and a considerable excess of silica, added in the form of 
coarse quartz grains, which dissolved rapidly in the melt 
above a gas burner. This mixture, held over night at 
about 1000°, gave well-developed crystals of the same 
form of SrO.Si0 2 . The identity, as in all the other cases, 
was ascertained by determination of the refractive 

Thus all the experiments failed to show any other form 
of SrO.Si0 2 than the one similar to the pseudowollas- 
tonite. 13 

The eutectic SrO.Si0 2 -2SrO.Si0 2 was located at 65.5 
per cent SrO, as is apparent from the diagram (fig. 1). 
The steep slope of the liquidus curves at both sides is 
remarkable. While in the 65 per cent SrO mixture the 
primary phase is SrO.Si0 2 , it is 2SrO.Si0 2 in the 67 per 
cent SrO mixture, and the liquidus has risen 60° above 
the eutectic temperature, 1545°. 

The existence of 3Sr0.2Si0 2 not proved. — The primary 
phase in all the mixtures between the eutectic at 65.5 per 
cent SrO and 2SrO.Si0 2 is strontium orthosilicate. The 
following experiments were carried out in order to es- 
tablish whether there might be formed the compound 
3Sr0.2Si0 2 , analogous to 3Ca0.2Si0 2 , known from the 
studies on the calcium silicates. 

Strontium carbonate and quartz powder in the propor- 
tion required to form 3Sr0.2Si0 2 (72 per cent SrO), well 
mixed together, were held 2 hours at 1475° and quenched. 
The product consisted of SrO.Si0 2 and 2SrO.Si0 2 . 
Another time the charge was allowed to cool slowly. The 
result was the same. 

The same mixture was heated 16 hours at 1150°. It 
was now so fine-grained and poorly individualized, that it 
could not be decided whether there were one or two 
phases present. Neither SrO.Si0 2 nor 2SrO.Si0 2 could 
be identified. 

13 Jaeger and Van Klooster (op. cit., p. 903) heated 0.5 g. of pure SrO.SiOo 
with 1 or 2 g. of sodium tungstate at 860° during 72 hours and always 
obtained the same form as by direct crystallization from a melt, whether 
they had started with crystals or with glass. 

P. EsTcola — Silicates of Strontium and Barium. 341 

Strontium ortho silicate, 2SrO.Si0 2 . — The melting 
point of 2SrO.SiQ 2 is far above the range of the platinum 
resistance furnace. 

In a search for several polymorphic forms of this com- 
pound I heated a preparation of crystalline 2SrO.Si0 2 at 
different temperatures: 5 minutes at 1634°, 2 hours at 
1415°, 1300°, and 990° respectively. The result was 
always the same form, well characterized by its refrin- 
o-ence and twinning. It was also obtained from a SrCl 2 
melt at about 1000° (cf. above, p. 339). Nor did the 
substance change when allowed to cool very slowly. The 
phenomenon of "dusting" in the case of 2CaO.Si0 2 
apparently has no analog here. 

For the determination of the optical properties the 
preparation quenched from 1634° was used, with the 

ring results. 




F ... 

. 1.740 



Tl .. 

. 1.7325 



D .. 

. 1.7275 



C ... 

. 1.722 



The crystals are optically positive with 2E = 58° 
approximately, or 2 V = 32° 30'. 

In many of the crystals a twinning was observed, often 
repeated (fig. 5). In sections normal to y, the acute 
bisectrix, the trace of the axial plane forms, with the 
composition planes of the twinning lamellae, angles of 17°. 
Thus, if the crystal system is monoclinic, the plane of the 
optic axes is normal to the plane of symmetry. 

The repeatedly twinned crystals, in their outlines, 
often display the habit of orthorhombic crystals, and 
when the lamellae are very narrow they have apparent 
straight extinction. 

14 Free strontium oxide attacks platinum on heating. Every mixture does 
so as soon as it contains an excess, however small, of SrO over 2SrO.Si0 2 , 
while the more acid mixtures may be heated in platinum crucibles without 
any danger. The phenomenon appears as a blackening of the surface of 
the platinum and of the charge. This black substance is soluble in HC1, 
forming chloroplatinic acid. 

A probable explanation is that the strontium oxide dissociates at high 
temperatures, either into metallic strontium or into a strontium suboxide 
which forms an alloy with platinum. This may be connected with the face 
that the strontium oxide is somewhat volatile. 

Both the dissociation and the volatilization are, however, very slight at the 
temperatures of the platinum resistance furnace. 0.300 g. SrO, wrapped 
in platinum foil and held 4 hours at 1580°, lost only 0.0017 g. in weight. 

34:2 P. Eskola — Silicates of Strontium and Ba 


Strontium oxide, SrO 1 *. — All the mixtures which 
contained more SrO than 2SrO.Si0 2 , resulted in two 
phases, 2SrO.Si0 2 and SrO, when heated at high 
temperatures. In these mixtures the strontium oxide 
could always be identified from its isotropic character 
and high index of refraction. 

Fig. 5, 




Fig. 5.— Crystals of 2SrO.Si0 2 . 

Pure SrO was prepared in two ways, from strontium 
carbonate and from strontium nitrate. In the first case 
it forms minute rounded, though clear grains. When 
these were held 4 hours at 1580° they reacted slowly with 
water and even with HC1. From nitrate the oxide may 
be obtained as large clear crystals showing a perfect 
cubic cleavage. 15 

The determination of the index of refraction was very 
difficult, because the oxide instantly reacts with the high 
refractive liquids containing iodine. It was found to be 
somewhat higher than 1.86 for all the colors and may be 
very roughly estimated at 1.87. 

15 See also G. Briigelmann, Z. anorg. Chem., 10, 415, 1895. 

P. Eskola — Silicates of Strontium and Barium. 343 

The specific gravity lias been determined by Briigel- 
mann (loc. cit.) as 4.750. 

According to H. Moissan, 16 strontium oxide melts more 
easily than calcinm oxide, but nothing more is known 
abont its melting temperature. 

Wyckoff 17 has determined, by Debye and Scherrer's 
method, the crystal structures of CaO, SrO, BaO and 
mixtures of CaO with SrO and BaO, respectively, in 
samples prepared by me. His results have interest for 
us in so far as they prove SrO to be perfectly isomor- 
phous with and to form a series of solid solutions with 

The System BaO-Si0 2 . 

The quenching experiments pertaining to the system 
BaO.SiOo are described in table II. 

Table II. 

Composition v 

rt. per 




es ! 



Temper- Time 


Si0 2 2Si0 2 

3Si0 2 

ature °C minutes Resulting phases 

Liquidus of tridymite 



1551 90 Glass and tridymite 



1472 120 Glass 



1447 90 Glass and tridymite 

Euteetie tridyniite-Ba0.2Si0 2 



1372 135 " Tridymite and Ba0.2SiO, 



1390 120 Glass 



1377 60 Glass and tridymite 



1373 120 Ba0.2SiO, and tridymite 



1398 120 Glass 



1385 60 Crystals and glass 



1369 120 Tridymite and Ba0.2SiO- 

Melting point of Ba0.2Si0 2 



1425 90 Glass 



1421 30 Glass 



1419 60 Ba0.2Si0 2 (with some gli 


of the solid solutions Ba0.2Si0 2 -2Ba0.3Si0 2 





1441 60 Glass 





1435 60 Glass 





1432 60 Glass and mix crystals 





1448 60 Glass 





1443 60 Glass and mix crystals 





1451 60 Glass 





1448 60 Glass and mix crystals 

IC H. Moissan, Ann. Chem. Phys., (7) 4, 136, 1895. 
17 Ealph W. G. Wyckoff, unpublished data. 

344 P. Eshola — Silicates of Strontium and Barium. 

Solidus of the solid solutions Ba0.2Si0 2 -2Ba0.3Si0 2 

58 42 71 29 1418 60 Crystals 

58 42 71 29 1424 60 Mix crystals with little gls 

60 40 42 58 1435 60 Mix crystals and glass 

60 40 42 58 1431 60 Mix crystals 

61.25 38.75 23.9 76.1 1443 60 Mix crystals and glass 

61.25 38.75 23.9 76.1 1440 60 Mix crystals and glass 

61.25 38.75 23.9 76.1 1438 60 Mix crystals 

Melting point of 2Ba0.3Si0 2 . 
62.90 37.10& 1451 60 . Glass 

62.90 37.10 1451 30 Glass 

62.90 37.10 1448 30 Glass and 2Ba0.3Si0 2 

Eutectic 2Ba0.3SiO,-BaO.Si0 2 
65 35 1439 15 Glass 

BaO.SiOo and 2Ba0.3Si0 2 
30 with a little glass. 

30 BaO.SiO, and 2Ba0.3SiO,. 

15 2BaO.SiO, and glass. 
15 2BaO.Si0 2 and BaO.SiO, 

From these determinations results the equilibrium 
diagram tig. 6. A part of it is given on a larger scale in 

The liquidus curves of cristobalite and tridymite in 
this binary system were, by means of a few quenchings, 
shown to have the same general character as in the 
system SrO-Si0 2 , and were not investigated any more 

The eutectic tridymite-BaO .2SiO 2 . — The mixture con- 
taining 47.5 per cent BaO was found to be very close 
to the eutectic composition, and on the tridymite side of 
it. The eutectic composition was placed at 47 per cent 
BaO, 53 per cent Si0 2 and the temperature at 1374°. 

Barium disilicate, Ba0.2Si0 2 , has previously been 
studied by N. L. Bowen. 18 Crystals of this composition 
were found in optical glass of the variety known as 
"barium crown." They had the shape of platy elon- 
gated hexagons (fig. 8a), about 3 mm long, 2 mm wide, 






Eutectic Ba( 







a — Ba0.2Si0 2 

b — 2Ba0.3SiO 

1S N. L. Bowen: Crystals of barium disilicate in optical glass. J. Wash. 
Acad. Sci., 8, 265, 1918. Cf. also N. L. Bowen: Devitrification of glass, 
J. Am. Ceram. Soc, 2, 261, 1919. 

P. Eskola — Silicates of Strontium and Barium. 345 


Fig. 6. 


- i 



1 1 1 

i ■ i i 










3 I 







~° \ 

•0 I 






VT) ' 


6 j 

ia / 






Crista b a lite 




•a / 

and liquid 


■a \ 

3 / 

or / 


3 \ 
O" 1 

■a / 






g MiXC 


rystals of r 







,£26a0-3SiO 2 -BaO55iO 5 / 
°-< and liquid / 














Tri dymite 



and liquid 







o ._ 

<n. »* 

— o 

(0 re 

+- CQ 

t \l 



Bao-2s;o 2 and 

1 iqui d 










— 6 




i i 

0^ and ~tridym'\te 

i i i 


2BaO-5iO« Ba05iO s 2Ba035i(fc Ba0-25i0 2 


Fig. 6. — Temperature-concentration diagram of the binary system 

and only 0.03 mm thick. The terminal angles were 
approximately 100° and the lateral angles 130°. The 
crystals are orthorhombic, the negative acute bisectrix 
being parallel to the elongation, and the plane of the optic 
axes parallel to the platy development, which is a good 
cleavage. 2V a = 70° approximately y = 1.613 ; a = 1.595. 

346 P. Eskola — Silicates of Strontium and Barium. 
In pure crystals obtained by cooling a melt of the compo- 

sition Ba0.2SiOo were found 

1.617 and a = 1.598. 

The melting point was determined by the heating curve 
method at 1426 °C. Bowen suggests that the slightly 
lower values of the indices of refraction of the crystals 
in the optical glass may represent a real difference and 
that when formed in the glass they take a small amount 
of alkaline disilicates into solid solution. 




BaO-4 « a *3Si0 2+ liquid 



" and \ I^-t 
Jiquid V/* 

\^~r^-^/ Mixer y st ais of 

1\>^\ SBaOaSiOi-BaOgSiOa 
>\^ ^\ +li<^uid 


Tridymite / " 
and liquidv/ " 


f\ - 

Ba0SlO £ + 

Mixcrystals of 



2BaO-35)0 A 


* N. / 



BaO 25io £ + 

t ridymite 

1 . 

67Wt%Ba0 65 



55.98 55 


54Wt°/oSi0 t 

Fig. 7. — Part of the binary system BaO-Si0 2 showing the melting of the 
mix crystals of barium disilicate and dibarium trisilieate. 

I prepared crystals of Ba0.2Si0 2 from a BaCl 2 -melt, 
and also obtained well-formed crystals embedded in 
Ba-silicate glass from a mixture containing 50 per cent 
BaO, held at 1385°. In both cases the crystals were 
bounded only by the three orthorhombic pinacoids, were 
very much elongated in the direction of « and had almost 
square cross sections (fig. 8b and 8c). There are notable 
cleavages in all the pinacoidal directions, but by far the 
most perfect is the one parallel to a/?, so that when 
crystalline masses of this substance, as for example, that 
obtained by cooling of the pure melt, are ground, there 
are formed almost exclusively scaly cleavage splinters 
normal to y. 

P. Eskola — Silicates of Strontium and Barium. 347 

Fig. 8. 
a. 1). c. 




Fig. 8. — Optical orientation in crystals of barium disilicate : a = crystals 
in optical glass after N. L. Bowen; ~b and c = crystals from BaCL-flux, and 
as formed in barium silicate melt. 

Thus these crystals are, in their habit, considerably 
different from those in the barium crown glass, as 
described by Bowen. In other ways my crystals agree 
well with the characteristics given by Bowen. 

The indices of refraction are as follows : 

o(D) =1.597; £(D) =.1.612; y(D) =1.621 

As appears from these values, also confirmed by direct 
observation, the optical character is negative and the 
axial angle approximately 75°. A calculation from the 
indices of refraction gives 2V = 74°45'. 

The density, determined with the pycnometer, was 
found to be 3.73. 

The melting point of pure Ba0.2Si0 2 , according to my 
measurements, is 1420 ± 4°. 

There was no indication that the barium disilicate takes 
silica in solid solution; the crystals obtained from the 
mixture containing 50 per cent Si0 2 showed exactly the 
same optical properties as those from the pure melt. On 
the other hand it forms a remarkable solid solution series 
with another less acid silicate. 

Am. Jour. Sci.— Fifth Series, Vol. IV, Xo. 23.— October, 1922. 

348 P. Eskola — Silicates of Strontium and Barium. 

Solid solutions of barium disilicate, Ba0.2Si0 2 and 
dibarium trisilicate, 2Ba0.3Si0 2 . — My first observations 
pointing' to the existence of solid solutions in the system 
Ba0-Si0 2 were made with the mixture containing 60 per 
cent BaO. It was found to give only one kind of crystals 
very similar but not identical with those of Ba0.2Si0 2 . 
Furthermore I found that the crystals, when formed in 
a solid state (below the solidus) had lower refractive 
indices than they had when formed in the presence of a 
liquid phase (above the solidus). These observations 
made it probable that there exists a solid solution series 
without a maximum or minimum. In the course of the 
work this proved to be the case, the other end member 
of the series being the compound 2Ba0.3Si0 2 . 

The gradual change in the physical properties will 
appear from the following description. 

Mixture 29 per cent 2Ba0.3Si0 2 -71 per cent Ba0.2Si0 2 
gives crystals which agree fairly well with the pure 
disilicate. Among the three pinacoidal cleavages the one 
parallel to a/3 is most perfect. The optical character, 
however, is distinctly positive, with 2V = 70° approxi- 

Density = 3.80. 

The crystals of the composition 58 per cent 2BaO. 
3Si0 2 -42 per cent Ba0.2Si0 2 show a different habit, 
being usually somewhat elongated along p. The cleavage 
parallel to Py is most perfect. Cleavage along a/? also is 
good and that parallel to ay distinct. Many crystals are 
composition twins along a prismatic face in the zone of a. 
The angle between p and the twinning plane is about 30°, 
or between p in both individuals about 60°. The optical 
character is positive and the axial angle quite small. I 
determined 2E = 44°, or 2V = 27°. 

Crystals of the composition 76.1 2Ba0.3Si0 2 -33.9 
Ba0.2Si0 2 are equant grains, not elongated. Cleavage 
along py is best, next to it that along a/3, while that 
parallel to ay was hardly observed at all, The same pris- 
matic twinning with the twinning plane parallel to a and 
inclined 30° to p was observed frequently, and is some- 
times poly synthetic. The axial character is positive and 
the axial angle larger than in the former. I measured 
2E = 56° or2V = 33°30'. 

The mix crystals containing 94.2 per cent 2Ba0.3Si0 2 
are in all their characters closely similar to the former 

P. Eskola — Silicates of Strontium and Barium. 349 

The variation of the refractive indices in this solid 
solution series appears from table III and from fig. 9. 

As may be seen from fig. 9, the indices of refraction, in 
the series of crystals between Ba0.2Si0 2 and 2BaO. 

Fig. 9. 









' ><; 




/ y 


'/ / / 




^ ^ 



_ -« ' 


__ __ _- ■ 




~~ • 

— - 1~"~~ / 



y . 


*^ > 



, ^ 


100 Ba0-2St0g 







Fig. 9. — Variation of the indices of refraction in the series of mix crystals 
of barium disilicate and dibarium trisilicate. * 

3Si0 2 , change gradually along with the composition, and 
all of them increase towards the latter end member, but 
their intermediate relations and hence the optic axial 
angles have rather complicated variations. 

350 P. Eskola — Silicates of Strontium and Barium. 

Table III. 

Wt. % 2Ba0.3Si0 2 29 58 76.1 94.2 100 

Wt. % Ba0.2Si0 2 100 71 42 23.9 5.8 

o(F) 1.602 1.602 1.619 1.622 1.625 1.627 

a(Tl) 1.599 1.599 1.615 1.619 1.622 1.623 

a(D) 1.597 1.597 1.612 1.616 1.619 1.620 

o(C) 1.595 1.595 1.610 1.613 1.616 1.617 

/3(F) 1.617 1.618 1.621 1.627 1.631 1.632 

j8(Tl) 1.614 1.614 1.618 1.623 1.627 1.628 

/3(D) 1.612 1.612 1.615 1.620 1.624 1.625 

/3(C) 1.610 1.610 1.612 1.617 1.621 1.622 

7 (F) 1.632 1.644 1.645 1.647 1.652 1.652 

7 (T1) 1.625 1.639 1.641 1.642 1.647 1.648 

7 (D) 1.621 1.636 1.638 1.639 1.644 1.645 

7 (C) 1.618 1.633 1.635 1.636 1.640 1.641 

Density^ 3.73 3.80 3.93 

a The determinations of density in this series were made on only 0.5 g. of 
substance and are therefore less accurate than in the other cases. 

It might seem probable, from the diagram, that a and 
p should exchange roles at abont 65 per cent 2Ba0.3Si0 2 . 
Drawn thus, the curves for p would be nearly straight 
lines. I have not, however, drawn the curves across each 
other because it was found that the twinning along a pris- 
matic face occurs in the same relation to the three main 
optical directions, in the two mixtures containing 58 and 
76 per cent 2Ba0.3Si0 2 . This indicates that the indices 
a and p have not exchanged roles. 

Dibarium trisilicate, 2Ba0.3Si0 2 - — This pure com- 
pound was observed only in the form of a granular mass 
and I did not see any crystalline forms. It does not seem 
to have any disposition for prismatic or scaly develop- 
ment, the grains being rounded and equant. The only 
cleavage that appears perfect is that along Py while the 
two other pinacoidal cleavages are poorly developed. 

The twinning and composition plane in the zone of a 
and inclined 30° from p forms polysynthetic lamellae. In 
many cases just a few such lamellae occur and they are so 
narrow that they appear as what is often called parting 

The indices of refraction (cf. table 3) are as follows: 

o(D) =1.620; /?(D) =1.625; y(D) =1.645. 

The optical character is positive and the axial angle is 
of about the same size as in diopside, or 2V = 58°. A 

P. Eskola — Silicates of Strontium and Barium. 351 

calculation from the indices of refraction gives 2V = 
53° 30'. The density was found to be 3.93. 

In summary, it appears that the two silicates BaO. 
2Si0 2 and 2Ba0.3Si0 2 are very perfectly isomorphous, 
have orthorhombic symmetry and pinacoidal cleavages. 
Their relations bear some resemblance to the plagioclase 
feldspar series. The melting diagrams of the solid solu- 
tion series are quite similar, cleavage variations are 
similar. The variation of the refractive indices and 
optic axial angles is irregular in the plagioclases, though 
in a smaller degree than we have found here. 

Very rarely do two compounds of the same element so 
different in chemical constitution form so perfect an 
isomorphous series. 

Eutectic 2Ba0.3Si0 2 -BaO.SiO ? .—The mixture of 60 
per cent BaO and 40 per cent silica was found to have 
very exactly the eutectic composition between the barium 
metasilicate and the dibarium trisilicate, as appears from 
the fact that a melt of this composition gives the two 
crystalline phases simultaneously. 

Barium metasilicate, BaO.Si0 2 . — In the BaO-Si0 2 
mixtures in wihch BaO.Si0 2 is a primary phase, this 
compound was found in the form of rounded globules or 
short rods with rounded ends. The latter show a nega- 
tive elongation and parallel extinction. In glasses 
composed of mixtures of barium and calcium silicates 
(in the system CaO.Si0 2 -BaO.Si0 2 ) it was frequently 
observed as small crystallites looking like short needles 
sharpened on both ends. These always show a parallel 
extinction and negative elongation. There is a well 
developed cleavage parallel to a/3. Grains placed parallel 
to this plane show the axial figures, with 2E = 50° 
approximately, or 2V = 29°. Dispersion of the optic 
axes is strong: p > v. Birefringence is very weak. 
Grains parallel to /3y y showing the trace of the good 
cleavage along af3, display abnormal blue interference 
colors. The indices of refraction were determined as 
follows : 




F ... 

.. 1.682 



Tl .. 

.. 1.677 



D .. 

.. 1.673 



C ... 

.. 1.669 



352 P. Eskola — Silicates of Strontium and Barium. 

The density at 4° was determined to be 4,399. 

The melting point of BaO.SKX is, according to Jaeger 
and Van Klooster, 1604 ±0.5°. 

To find ont whether the barinm metasilicate conld be 
inverted into some other form I heated a sample of the 
crystallized compound, first alone and a second time with 
one-fifth of its weight Ba(V0 3 ) 2 , over night at about 
1100°. No change took place. 19 

Eutectic BaO.Si0 2 -2BaO.Si0 2 . — A mixture of these 
compounds, containing 75 per cent BaO, gave at 1546° 
crystals of both kinds, but at 1556° only crystals of 2BaO. 
Si0 2 and glass. From this it may be concluded that 
there is a eutectic at about 74.5 per cent BaO with a melt- 
ing point of 1551 ±6°. 

No other compounds except the two named above occur 
in this part of the system. 

Barium ortho silicate, 2BaO.Si0 2 . — Barium orthosili- 
cate appears as a granular mass of rounded grains and 
does not show any cleavages or twinning. The melting 
point is higher than that of platinum. 

a (D) = 1.810 ± 0.005; y(D) =1.830 ± 0.005. 

Barium oxide, BaO. — Mixtures of barium orthosilicate 
and barium oxide were not studied more closely. The 
fact that all the mixtures containing an excess of BaO 
over the orthosilicate ratio attack platinum, seems to 
prove that they contain free BaO. 20 

Coarsely crystalline barium oxide was prepared from 
barium nitrate by heating slowly in a graphite crucible. 
It was obtained in clear translucent crystals showing 
cubic cleavage. 

The refractive index is very high, but difficult to deter- 
mine; the substance absorbs moisture from the air 
almost immediately when exposed and also attacks the 
high-refractive media (in this case mixtures of sulphur 
and selenium). I found (in red light) : 

n = 2.16 ± 0.05. 

Briigelmann has determined the specific gravity of 
barium oxide as 5.722. 

19 Jaeger and Van Klooster (loc. cit.) were not more successful in their 
attempt to effect a change by heating 0.5 g. barium metasilicate with 1 g. 
sodium tungstate at 860° for 72 hours. 

20 Barium oxide attacks the platinum in the same way as strontium oxide, 
but still more strongly. 

P. EsJcola — Silicates' of Strontium and Barium. 35& 

The System CaO.Si0 2 -SrO.Si0 2 

The quenches made to establish the melting diagram of 
mixtures of the metasilicates of calcium and strontium 
are recorded in table IV. 

Table IV. 

Composition wt. per cent 



Silicates Temper 



SrO. ature 


Kesulting phases 



Si0 2 

Si0 2 

Si0 2 °C 






25 1495 







25 1485 


Mix crystals 





37.5 1482 


Mix crystals and glass 





37.5 1474 


Mix crystals 





50 1482 







50 1478 


Glass and mix crystals 





50 1474 


Mix crystals 





56.25 1477 







56.25 1472 


Mix crystals 





75 1511 







75 1507 


Glass and mix crystals 





75 1500 


Mix crystals and glass 





75 1496 


Mix crystals 

As these data, together with the optical study of the 
mix crystals obtained, seemed fully to establish the 
general character of the melting diagram (fig. 10), it was 
considered unnecessary to determine the curve in more 

Fig. 10. 

CaO-SiO^ SrOSiO^ 

Fig. 10. — Temperature-concentration diagram of the binary system 
CaO.Si0 2 -SrO.Si0 2 showing a complete series of mix crystals with a minimum. 

detail. Assuming as the melting point 21 of a-CaO.Si0 2 , 
1540 ° ; and of SrO.Si0 2 , 1578° (cf. above, p. 335), we get 
the diagram reproduced in fig. 10. 

21 G. A. Eankin and F. E. Wright, this Journal (4) 39, 1, 1915. 

354 P. Eskola — Silicates of Strontium and Barium. 

The minimum melting temperature in this series was 
found to be 1474 ± 3° at the composition 44 per cent 
CaO.SiO ? -56 per cent SrO.Si0 2 . 

The mix crystals were found to be crystallographically 
very like a-CaO.Si0 2 . Considerable care was taken to 
find out whether they show that twinning on the base 
which, in the case of pseudowollastonite, indicates mono- 
clinic symmetry. 

In a-CaO.Si0 2 I found frequently poly synthetic twin- 
ning along the basal plane, as described by Wright (loc. 
cit.). In mix crystals with 12.5 per cent SrO.Si0 2 the 
twinning was no less clear; in one case I measured an 
extinction angle of 3°. In the mixture with 25 per cent 
SrO.Si0 2 twinning was clearly observed, while, in 
mixtures with 37.5 per cent, 56.25 per cent and 62.5 per 
cent SrO the extinction angles were apparently so small 
that the twinning lamellae, though occasionally positively 
identified, always appeared extremely faint. In a 
preparation with 75 per cent SrO.Si0 2 the occurrence of 
faint twinning along the base was still clearly observed. 
In the pure SrO.Si0 2 I did not succeed in finding definite 
twinning appearing as oblique extinction, in spite of 
much search. 

In summary, therefore, it may be stated that all the 
crystals containing calcium and strontium metasilicates 
in solid solutions are pseudohexagonal and really mono- 
clinic. In the pure strontium metasilicate, although it 
may belong to the same class of symmetry as the mix 
crystals, no deviation from hexagonal symmetry was 
observed (cf. p. 337). 

The indices of refraction for the solid solution series 
were determined with the following results. 

Table V. 

Wt. % CaO.SiO, 






Wt. % SrO.Si0 2 








































7 <D) 










P. Eskola — Silicates of Strontium and Barium. 355 

The same is expressed in the diagrams (fig. 11) in 
which is also shown the variation in the index of refrac- 
tion in the strontium and calcium metasilicate glass. 




Fig. 11. — Variation of the indices of refraction in the mix crystal series 
CaO.Si0 2 -SrO.Si0 2 . 

The System CaO.Si0 2 -BaO.Si0 2 . 

While strontium and calcium metasilicates give • a 
complete series of solid solutions and apparently are 
perfectly isomorphous, barium metasilicate does not mix 
at all with calcium metasilicate. Instead of this there 
occurs a double compound, 2CaO.Ba0.3Si0 2 , which, how- 
ever, has no true melting point, but breaks np into a-CaO. 
Si0 2 and liquid. Those quenches which yielded impor- 
tant results are quoted below in table VI. 

Fig. 11. 




P Tl 



— -^* 



> — — 


__ _s - 

— •- 


' y~ 

in D 

glasses c 

K - — 

r - 

— — •- 

— • 
_ «_ — 



_ -—-^ 









• ■ 

356 P. Eskola — Silicates of Strontium and Barium. 

Table VI. 

Composition wt. per cent 

Oxides Silicates Temper- Time 

CaO. Ba0 2 ature min- Besulting Phases 

CaO BaO Si0 2 Si0 2 Si0 2 °C utes 

Liquidus of a-CaO.Si0 2 
36.13 17.95 45.92 75 25 1466. 15 Glass 








Glass and a-CaO.Si0 2 








Glass and a-CaO.Si0 2 








a-CaSi0 3 and glass 
















Glass and a-CaO.Si0 2 
















Glass and a-CaO.Si0 2 








Glass and a-CaO.Si0 2 








Glass and a-CaO.Si0 2 




2CaO.Ba0.3Si0 2 -liquid. 








Glass and a-CaO.Si0 2 








Glass and 2CaO.Ba0.3Si0 2 








Glass and 2CaO.Ba0.3Si0 2 








Glass and 2CaO.Ba0.3Si0 2 
















Glass and 2CaO.Ba0.3Si0 2 

22.88 37.69 39.43 47.5 52.5 1314 15 Glass and 2CaO.Ba0.3Si0 2 

Decomposition of 2CaO.Ba0.3Si0 2 

25.13 34.32 40.54 52.18 47.82 1320 30 a -CaO.Si0 2 and glass 

= 2CaO.BaO.3SiO, 

25.13 34.32 40.54 52.18 47.82 1315 20 2CaO.Ba0.3Si0 2 only 

Liquidus of 2CaO.Ba0.3Si0 2 . 

21.68 39.48 38.84 45 55 1321 30 Glass 

21.68 39.48 38.84 45 55 1316 30 Glass and 2CaO.Ba0.3Si0 2 

21.68 39.48 38.84 45 55 1301 30 Glass and 2CaO.Ba0.3Si0 2 

16.99 46.47 36.54 35.26 64.741306 20 Glass 

(BaO.Ca0.2Si0 2 ) 

16.99 46.47 36.54 35.26 64.741300 20 Glass and 2CaO.Ba0.3Si0 2 

16.99 46.47 36.54 35.26 64.74 1294 20 Glass and 2CaO.Ba0.3Si0 2 

16.38 47.37 36.25 34 66 1293 20 Glass and 2CaO.Ba0.3Si0 2 

16.38 47.37 36.25 34 66 1290 20 Glass and 2CaO.Ba0.3Si0 2 

15.66 48.45 35.89 32.5 67.5 1289 15 Glass and 2CaO.Ba0.3Si0 2 

15.42 48.81 35.77 32 68 1290 15 Glass 

15.42 48.81 35.77 32 68 1284 25 Glass and 2CaO.Ba0.3Si0 2 

14.45 50.25 35.30 30 70 1280 20 Glass 

14.45 50.25 35.30 30 70 1276 20 Glass and 2CaO.Ba0.3Si0 2 

14.45 50.25 35.30 30 70 1274 20 Glass and 2CaO.Ba0.3Si0 2 

Eutectic 2CaO.Ba0.3Si0 2 -BaO.SiO a 

16.99 46.47 36.54 35.26 64.74 1273 2CaO.Ba0.3Si0 2 and glass 

16.99 46.47 36.54 35.26 64.74 1271 2CaO.Ba0.3Si0 2 and glass 

16.99 46.47 36.54 35.26 64.74 1265 2CaO.Ba0.3Si0 2 and BaO. 

Si0 2 

P. Eskola — Silicates of Strontium and Barium. 357 







2CaO.BaO.3SiO, and glass 







2CaO.Ba0.3Si0 2 and BaO. 
Si0 2 





















2CaO.BaO.3SiO, and BaO. 
Si0 2 







BaO.SiO, and glass 







BaO.SiO, and 2CaO.BaO. 







BaO.SiO, and 2CaO.BaO. 

. Liquidus of BaO.SiO, 

12.05 53.83 34.12 25 75 1300 Glass 

12.05 53.83 34.12 25 75 1292 Glass and BaO.SiO, 

9.63 57.42 32.95 20 80 1367 Glass and BaO.SiO, 

The melting diagram resulting from these facts is 
given in fig. 12. 

The liquidus curve of a-CaO.Si0 2 was followed to 50 per 
cent BaO.Si0 2 . The a-CaO.Si0 2 which always separated 
in the form of thin crystals tabular parallel to the basal 
plane did not seem to take any BaO.Si0 2 in solid solution, 
as the crystals in all cases showed the indices of refrac- 
tion characteristic of pure pseudowollastonite : <*(D) = 
1.609 ± 0.002; y(D) =1.652 ± 0.002. 

Dicalcium barium silicate, 2CaO.Ba0.3Si0 2 . — When 
the mixture composed of 50 per cent BaO.Si0 2 and 50 per 
cent CaO.Si0 2 was allowed to crystallize on cooling, it 
formed a coarsely crystalline, fibrous mass, almost like 
natural wollastonite. The fibers were speckled with 
minute crystals of another substance (BaO.Si0 2 ). 
Supposing the fibrous crystals to have the composition 
2CaO.Ba0.3Si0 2 I prepared such a mixture, which was 
now found to crystallize as homogeneous crystals and on 
heating to break up at 1320 ± 4° into a-CaSi0 2 and 

The crystals are uniaxial and negative, probably hex- 
agonal, and have good cleavages in their prismatic zone. 
A crystallizing mass develops negatively elongated 

The indices of refraction were determined as follows : 



F ... 

;". 1.690 


Tl ... 

.. 1.685 


D ... 

.. 1.681 


C ... 

.. 1.677 


358 P. Eshola — Silicates of Strontium and Barium. 

Fig. 12. 

1 640 

1 1 1 1 — 

— i 1 1 1 1 

1 600 






/ " 



/ - 







oc-CaO-SiOs. \ 


and 1 (' c( u \ d' \ 




x / : 

1 ' ^~-^ / BaO-SiOj. 

^\ t and liquid 
£CaOBaO-35i0 2 :C \ r 


ot-CaO-S'i 0^ 

and liquid *\/- 

+ * + + + 



and 2 CaO.BaO-35iO a 



2CaO- Ba0-35i0 2 and BaO-5iO a 

1 200 



/3-CaO-Sio 5 and 2Ca0-Ba0-3 S«O a 

10 50 30 40 50 60 70 80 SO 100 

Ca0-Si0 a _ BaOSiOa. 

Tig. 12. — Temperature-concentration diagram of the binary system 
CaO.Si0 2 -BaO.SiCK 

As it was thought possible that there could exist solid 
solutions between the double compound and BaO.Si0 2 , 
I determined the refractive indices of the exceedingly 
small crystals in preparations with 55 and 60 per cent 

P. Eskola — Silicates of Strontium and Barium. 359 

BaO.SiO,. I found in both cases «(D) = 1.683 ± 0.003; 
e(D) = 1.669 ± 0.003. The results are not significantly 
different from those of the pure compound. 

The invariant point CaO.Si0 2 -2CaO.Ba0.3Si0 2 -melt. — 
In a mixture containing 50 per cent of CaO.Si0 2 and as 
much BaO.Si0 2 the crystals of 2CaO.Ba0.3Si0 2 break up 
into a-CaO.Si0 2 and liquid before melting entirely. In 
another mixture with 47.5 per cent CaO.Si0 2 the last 
crystals before complete melting consist of the double 
compound. The invariant point therefore is between 
these limits, and the course of both liquidus curves meet- 
ing there indicates it to be at 52 per cent BaO.Si0 2 and 
•48 per cent CaG.Si0 2 , with the temperature of complete 
melting 1320 ± 4°. 

The liquidus of 2CaO.Ba0.3Si0 2 has, from the 
invariant point, a regular course towards a eutectic. I 
followed its course very closely, because it was suspected 
that another double compound, CaO.Ba0.2Si0 2 , might 
possibly form here, its composition corresponding to 
35.26 weight per cent CaO.Si0 2 and 64.74 weight per cent 
BaO.Si0 2 . But no such compound could be isolated, and 
the corresponding mixture crystallized within an interval 
of 32 : into a verv fine mixture of BaO.SiOo and 2CaO. 
Ba0.3Si0 2 . 

The liquidus of BaO.Si0 2 was determined, with only a 
few quenches, to have a steep course towards the high 
melting point of the barium metasilicate. 

The crystals of this compound in melts containing 75 to 
80 per cent BaO.Si0 2 were not equant grains as in the 
system BaO.Si0 2 , but needle-like, sharp at both ends. 
As they have formed at much' lower temperatures than 
the crystals in BaO-SiO? melts, it may be possible that 
they represent another form of BaO.Si0 2 . The refrac- 
tive indices, however, so far as they could be determined, 
were identical, and experiments carried out to establish 
the transformation expected failed, even when minerali- 
zers were applied (cf. p. 352). 

The Properties of the Glasses. 

The physical properties of representative glasses 
studied in the present work are tabulated below. 

360 P. Eskola — Silicates of Strontium and Barium. 

Table VII 



in the system 

SrO-Si0 2 . 

Wt. % SrO 
Wt.% SiO, 



Sr0.2Si0 2 




SrO.Si0 2 










a Calculated from the value of specific gravity, <Z(4°) =3.540, given by 
Jaeger and Van Klooster, op. cit. 902. 

Table VIII. 

Glasses in the system BaO-Si0 2 . 

"Wt. % BaO 


Wt. % Si0 2 
















Ba0.2Si0 2 


















L3, 14. 






1.6 60 











































O 5 










eoWU-SiO^ 50 40 30 60WteSi0£ 50 40 

Fig. 13. — Variation of the indices of refraction in strontium silicate 

Fig. 14.t — Variation of the indices of refraction in barium silicate glasses. 


P. Eskola — Silicates of Strontium and Barium. 361 

Figs. 13 and 14 are diagrams of the variations in the 
refractive indices of glasses of the systems SrO Si0 2 and 
BaO-Si0 2 showing also the dispersions. 

Fig. 15. 








.5 50 



Si0 2 

Fig. 15. — Variation of the index of refraction in strontium and barium 
silicate glasses. 

Fig. 15 represents the same for the sodinm light. The 
curves are drawn as far as the point corresponding to the 
index of refraction of silica glass and extrapolated up to 
the points of the hypothetical refractive indices of 
barium and strontium oxide glasses. In this part the 
curves, of course, can not give the values of the indices 

362 P. Eskola — Silicates of Strontium and Barium. 

very exactly, while those of the acid end should be fairly 
accurate, and the data given here may be, in the case of 
the barium silicate, of some practical interest to the 
manufacturer of optical glasses. 

In tables IX and X are given the determinations of 
refractive indices of glasses in the systems CaO.Si0 2 - 

SrO.Si0 2 and CaO.SiO,-BaO.SiO ? 
expressed in fig. 16. 


same is 

Fig. 16. 

100 wt % 80 

Fig. 16. — Variation of the' indices of refraction in glasses of CaO.Si0 2 . 
SrO.Si0 2 and CaO.Si0 2 -BaO.Si0 2 . 

P. Eskola — Silicates of Strontium and Barium. 363 



Glasses in the 

system CaO.Si0 2 - 

-SrO.Si0 2 . 

;.% CaO.SiO, 






;. % SrO.SiO., 


































Table X. 

Glasses in the system CaO.Si0 2 -BaO.Si0 2 . 

Wt. % CaO.SiO, 100 75 60 50 40 25 

Wt. % BaO.SiO, 25 40 50 60 75 100a 

?i(F) 1.635 1.643 1.649 1.6525 1.658 1.6655 (1.681) 

n(Tl) 1.631 1.638 1.6435 1.647 1.653 1.6605 (1.675) 

n(D) 1.628 1.6345 1.6395 1.644 1.649 1.657 (1.672) 

?i(C) 1.625 1.6305 1.636 1.640 1.645 1.653 (1.668) 

Density 3.633 

a Extrapolated. 

Absence of Diopside Analogs. 

A few experiments were made to ascertain whether 
strontium and barium metasilicates form double com- 
pounds with magnesium metasilicate, analogous to diop- 
side, the calcium compound, CaO.Mg0.2Si0 2 . Mixtures 
corresponding to these proportions were prepared and 
quenched from different temperatures. 

A mixture corresponding to SrO.Mg0.2Si0 2 gave only 
glass at temperatures above 1320 and, at temperatures 
down to about 1200°, glass and crystals that from their 
optical properties may have consisted of clinoenstatite. 
Another mixture of the composition 2SrO.Mg0.3Si0 2 
gave SrO.Si0 2 as a primary phase. It therefore seems 
that the metasilicates of strontium and magnesium do not 
form any double compound at all. At any rate, there is 
no compound SrO.Mg0.2Si0 2 , analogous to diopside. 

Furthermore, I made a mixture of 25 weight per cent 
SrO.Mg0.2Si0 2 and 75 weight per cent CaO.Mg0.2Si0 2 , 
heated it above 1400° and let it cool slowly. The product 
contained diopside, showing a(D) = 1.665 ± 0.002; 
/?(D) =1.675 ± 0.003; y(D) =1.695 ± 0.002 and 2V = 
60°. These properties agree perfectly with those of 
pure diopside and consequently the crystals do not 
contain any strontium compound as isomorphous mix- 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 23. — November, 1922. 

364 P. Eskola — Silicates of Strontium and Barium. 

ture. This conclusion, however, is not very dependable, 
as our experience is that strontium replacing calcium 
changes the optical properties of crystals very little. 
More weighty evidence proving positively that the 
amount of the strontium compound entering into solid 
solution in the diopside must be very small, is that the 
product consisted of more than one phase, the other 
phases besides the diopside being present as a fine 
crystalline mass. 

The preparation of the composition BaO.Mg0.2Si0 2 , 
on cooling from a melt heated not above 1370°, also gave 
several phases, and no double compound analogous to 
diopside was formed. 

This negative result is very remarkable, especially in 
the case of strontium, which in other ways showed such 
close similarity with calcium. 

The Strontium and Barium Feldspars. 

Former investigations. — From melts of the corre- 
sponding oxides Fouque and Michel-Levy 22 believed they 
had prepared strontium, barium, and lead analogs of 
anorthite, oligoclase and labradorite. 

Their different "feldspars" had the following specific 
gravities : 

Sr Ba Pb 

Oligoclase 1.619 2.906 3.196 

Labradorite 2.862 3.333 3.609 

Anorthite 3.043 3.573 4.093 

The products were not well crystallized and their crystal 
system could not be determined with certainty. The 
barium feldspar, BaAl 2 Si 2 8 , always formed microlites 
showing parallel extinction with negative elongation, and 
the crystals, being usually rectangular in cross section, 
seemed to be orthorhombic. Later investigation by E. 
Dittler, 23 on artificial barium feldspar, confirmed this 
result and also established it to be biaxial. Ginsberg, 24 

B2 F. Fouque et A. Michel Levy : Sur la production artificielle de f eldspaths 
a base de baryte, de strontia et de plomb, correspondant a 1 'oligoclase, 
a labradore et a 1 'anorthite; etudes des proprietes optiques des ces miner - 
aux, Bull. soc. min. France 3, 124. 1880. 

23 E. Dittler, Tscherm. Min. Petr. Mitt,, 30, 122, 1911. 

24 A. S. Ginsberg, Ann. de 1 'Inst. Polyteeh. Pierre le Grand a, Petrograde, 
1915, XXIII. 

P. Eskola — Silicates of Strontium and Barium. 365 

however, states that crystals of BaAl 2 Si 2 O s obtained by 
cooling of the pure melt were uniaxial, positive, and he 
regards this form as presenting a nephelite analog. 

There is some evidence to the effect that some amount 
of lime, in plagioclase feldspars, may be replaced by 
baryta without changing the triclinic symmetry. Des- 
cloizeaux 25 found this to be the case in a natural feldspar 
from unknown locality, whose optical and crystallo- 
graphic properties were similar to those of labradorite, 
while its composition resembled that of oligoclase more 
closely, with 7.30 BaO, 1.83 CaO, 7.45 Na 2 0, and 0.83 K 2 0. 
Ginsberg 26 investigated the binary system CaAl 2 Si 2 8 - 
BaAl 2 Si 2 8 in artificial products and found that the tri- 
clinic anorthite may take up limited amounts of barium 
feldspar in solid solution, whereby the optic axial angle 
diminishes. The hexagonal BaAl 2 Si 2 8 , on the other 
hand, may take up some lime feldspar. 27 

The relation of barium feldspar and orthoclase is far 
better known. Penfield 2S described barium-bearing 
orthoclase from Blue Hill, Delaware County, Pennsyl- 
vania. Later Sjogren 29 and Strandmark 30 studied the 
barium feldspar, or celsian, from Jakobsberg, Sweden, 
and proved it to have the composition BaO.Al 2 3 .2Si0 2 . 
The latter establishes its monoclinic symmetry and 
crystallographic similarity to adularia. The crystals are 
usuallv elongated parallel too. The angle a A o = 3°1' 
in the acute angle p. <* = 1.5835 ; p = 1.5886 ; y = 1.5941. 
Sp. g. 3.37. Strandmark also proved that the hyalo- 
phanes are isomorphous mixtures of celsian and ortho- 

Rock analyses including determinations of BaO and 
SrO commonly show small amounts of both of these 
oxides. As Washington 31 has pointed out, the highest 
amount of both of these oxides have been found in certain 
highly potassic rocks of Wyoming, showing up to 1.10 
per cent BaO and 0.30 per cent SrO. A study of the 
analyses published in Washington's Tables 32 shows that 

23 A. Descloizeaux, Tscherm. Min. Petr. Mitt., 7, 99, 1877. 

26 A. S. Ginsberg, loc. cit. 

27 The thermal work of Ginsberg was done by the cooling curve method 
and is subieet to criticism (cf. p. 332). 

28 S. L. Penfield, this Journal, 36, 326, 1888. 

29 Hj. Sjogren, Geol. Foren., Forh., 17, 578, 1895. 

30 J. F. Strandmark, Geol. Foren. Forh. 25, 289, 1903 ; 26, 97, 1903. 

31 H. S. Washington, J. Franklin Inst., 190, 767, 1920. 

32 H. S. Washington, U. S. Geol. Surv., Prof. Paper 99, 1917. 

366 P. Eskola — Silicates of Strontium and Barium. 

the ratio of the percentages of BaO and SrO corresponds, 
to a considerable degree, with the ratio of the percentages 
of K 2 and CaO, so that it may be presumed that strontia 
is present in the plagioclase and baryta in the potash 
feldspar. A chemical investigation of the feldspars 
from the rocks in question might prove this to be true. 

In the present study it was not intended to make any 
exhaustive experimental investigation of the strontium 
and barium feldspars, but a few experiments were made 
to elucidate their behavior. 

As both the compounds, SrO.Al 2 3 .2Si0 2 and BaO. 
Al 2 3 .2Si0 2 , were found to have very high melting 
points (far above 1700°) and to crystallize but poorly 
when the pure mixtures of the respective oxides were 
heated up, I applied the vanadate flux method to produce 
better crystals. In both cases the feldspar mixtures 
were powdered with about 1/3 of their masses of 
Sr(V0 3 ) 2 and Ba(V0 3 ) 2 , respectively, and kept over 
night at about 1400°. The vanadates were then washed 
out with very dilute cold HC1. 

Artificial barium feldspar, BaO.Al 2 3 .2Si0 2 . — The 
barium feldspar thus obtained formed minute crystals, 
not more than 0.05 mm in length. They were elongated 
parallel to a and tabular parallel to ay, so that most 
crystals placed themselves on that plane and, in conver- 
gent light, showed the trace of the optical normal. The 
terminal faces show an apparent bilateral symmetry, like 
orthorhombic domes and, as the extinction is almost 
always parallel so far as can be determined, the crystals 
give entirely the appearance of orthorhombic symmetry, 
in agreement with the earlier results of Fouque and 
Michel-Levy, and of Dittler. I found, however, that this 
is only apparent and due to the fact that most of the 
crystals are Carlsbad contact twins, so that light has to 
pass through the two individuals with opposite extinction 
angles, the individuals being tabular parallel to 010, 
which is also the composition plane. The other forms 
are (110), which is the zone of elongation, and (001) and 
(101). In simple crystals, or in outstanding simple 
parts of the twins, there may be observed extinction 
angles of two or three degrees, and the optical orientation 
is thus: || b; a A c = 3°. Basal cleavage is not very 
well developed, though its trace may be seen occasionally. 
All these characters agree with those found in natural 

P. Eskola — Silicates of Strontium and Barium. 367 

celsian. No evidence was found of the existence of a 
nephelite analog, a hexagonal form of BaAl 2 Si 2 O s 
described by Ginsberg. 

The indices of refraction indicate a negative character. 
They were determined as follows and are compared with 
Strandmark's result on the natural celsian: 

a(D) /3(D) 7 (D) 

Artificial barium feldspar' 3 ...... 1.587±0.002 1.593±0.002 1.600±0.002 

Natural celsian 1.5835 1.5886 1.5941 

The differences in the indices of refraction are some- 
what greater than the probable errors, but not more than 
may be accounted for by the fact that the natural celsian 
carries a few tenths of a per cent of alkalies and lime. 

Artificial strontium feldspar, Sr0.Al 2 O 3 .2SiO 2 . — The 
strontium feldspar obtained from the vanadate flux 
formed a crystalline mass showing radiating fibrous 
development but no well-formed crystals. No twinning 
could be discerned. Therefore nothing can be said about 
its crystallographic characters. The indices of refrac- 
tion, however, were readily determined and are stated 
below, compared with those of anorthite. 

a(D) /3(D) 7 (D) 

SrO.Al,0 3 .2SiO, 1.574±0.002 1.582±0.002 1.586±0.002 

CaO.AlA.2SiO, ..1.576 1.584 1.588 

It was thus found that the strontium feldspar is, in its 
optical properties, exactly like the calcium feldspar 
within the limits of the possible errors. This is not very 
surprising in itself, as we have found in some cases the 
indices of refraction of the strontium compounds some- 
what lower and in other cases somewhat higher than 
those of the calcium compounds. 

A mixture of 50 weight per cent SrO.Al 2 3 .2Si0 2 and 
50 weight per cent CaO.Al 2 3 .2Si0 2 was prepared and 
heated at about 1500°. It formed clear homogeneous- 
looking grains whose indices of refraction agreed with 
those of anorthite. It did not melt at this temperature 
and it is therefore probable that the strontium feldspar 
forms a complete series of solid solutions with anorthite. 

33 a (F) =1.593; a (Tl) = 1.590; a (D) =1.587; a(G) =1.585 

368 P. Eskola — Silicates of Strontium and Barium. 

Some General Considerations Regarding the Relations of the 
Alkaline Earth Compounds. 

"With the exception of some rare titanosilicates, the 
only natural anhydrous silicates of barium known are 
barium feldspar, or celsian, BaAl 2 Si 2 8 , notable for its 
isomorphism with potash feldspar, barylite, Ba 4 Al 4 Si 7 - 
24 , and taramellite, Ba 4 Fe"Fe" / 4 Si 10 O 31 . Hydrated 
silicates are more numerous, mostly belonging to the 
zeolite group, namely: brewsterite, H 4 (Sr,Ba,Ca)Al 2 Si 6 - 
1? + 3H 2 0, harmotome, H 2 (K 2 ,Ba)Al 2 Si 5 15 + 4H 2 0, 
edingtonite, BaAl 2 Sio0 10 + 3H 2 and wellsite (Ba,Sr,Ca,- 
K 2 )Al 2 Si 3 O 10 + 3H 2 0. Micas sometimes contain appre- 
ciable amounts of barium. 

No strontium silicates are known as minerals, though 
this element, in smaller quantities, is contained in some 
rock-forming feldspars, probably plagioclases (cf. p. 365), 
and in hancockite, a lead-epidote, and in zeolite minerals 
heulandite, H 4 (Ca,Sr)Al 2 Si 6 18 + 3H 2 0, and wellsite 
and brewsterite, named above. 

The present study adds some information concerning 
the isomorphous relations of a number of other silicates 
of strontium and barium. Thus the compound BaO. 
2Si0 2 which was suggested by Bowen (loc. cit.) to be 
isomorphous with the potassium disilicate studied by 
Morey and Fenner, 34 has now been found to be also iso- 
morphous with 2Ba0.3Si0 2 . A strontium metasilicate 
was found to be isomorphous with a form of calcium 
metasilicate, and a strontium orthosilicate is probably 
isomorphous with either the a or the p form of calcium 
orthosilicate. A strontium feldspar was found to be 
probably isomorphous with the lime feldspar, anorthite, 
which is also isomorphous with a soda feldspar. 

Now taking into account all the compounds of the alka- 
line earths, silicates as well as others, we may discrimi- 
nate, according to their isomorphous relations to com- 
pounds of other elements, at least four different classes : 

(1) Calcium, strontium, and barium compounds iso- 
morphous with each other (all or two of them) and also 
isomorphous with lead compounds. Examples : The 
tetrahedral-pentagondodecahedral nitrates of Ca(?), Sr, 
Ba, and Pb. The orthorhombic carbonates, aragonite, 
strontianite, witherite, and cerussite. The sulphates, 

34 G. W. Morey and C. N. Fenner, J. Am. Chem. Soc, 36, 215, 1914. 

P. Eskola — Silicates of Strontium and Barium. 369 

celestite, barite, and anglesite, form an isomorphous 
series which excludes calcium sulphate. Among silicates 
we note the group of hancockite and epidote, and we have 
now found that the metasilicates form an isomorphous 
series to which belong only the calcium and strontium, 
but not the barium metasilicate. 35 

(2) Calcium and strontium (also barium!) com- 
pounds isomorphous with each other and also isomor- 
phous with sodium compounds. Example: Anorthite 
and strontium anorthite. 

(3) Barium compounds isomorphous with potassium 
compounds. Example : Barium disilicate. Among the 
natural silicates named above, celsian and harmotome 
are representatives of this class. 

(4) Calcium compounds isomorphous with corre- 
sponding compounds of magnesium, ferrous iron and a 
number of other elements, but not with those of barium or 
strontium. Example : The rhombohedral carbonates, 
calcite, magnesite, siderite, etc. 

It seems to be a rule that in the compounds in which 
the lime may possibly be replaced by magnesia and 
ferrous oxide, it can not be replaced by strontia or baryta. 
Thus it was found in the present work that there are no 
strontium or barium compounds analogous to diopside, 
CaO.Mg0.2Si0 2 . 

Accordingly we find, in nature, strontium and barium 
compounds forming mix crystals with lime and alkali 
minerals, but not with ferromagnesian, although the 
latter may contain calcium. In other words, these 
elements are likely to be found in salic rather than in 
femic rocks, a circumstance that is really very striking 
in their distribution in the igneous rocks. 

The authors of the so-called quantitative classification 
of igneous rocks make a distinction between salic and 
femic lime. Now these terms gain added significance, as 
we find that only the salic but not the femic lime may be 
replaced by strontia. 

It may be of interest, for the sake of comparison, to 
present together some important properties of the 
members of some well known simple compounds of the 
three alkaline earth metals. 

33 Bourgeois (Ann. chim. phys., 29, 445, 1883) records a lead metasilicate 
of a similar appearance to those of calcium and strontium. 

370 P. Eskola — Silicates of Strontium and Barium. 

■ Mol.vol. ( a +p+ y )/3 

Carbonates : Aragonite 34.01 1.632 

Strontianite 39.87 1.615 

Witherite 45.82 1.627 

Sulphates : Anhydrite 45.99 1.586 

Celestite 46.18 1.626 

Barite 51.96 1.641 

Metasilicates : £-CaO.Si0 2 40.09 36 1.625 

«-CaO.Si0 2 40.11 36 1.625 

SrO.Si0 2 44.91 1.612 

BaO.Si0 2 48.57 1.675 

The carbonate series exemplifies the rule found to hold 
good in many other instances that the molecular volumes, 
in an isomorphous series, increase regularly with the 
atomic weights of the substituted elements. Compounds 
which are not isomorphous often show a discrepancy in 
their molecular volumes, as for example anhydrite in its 
relation to celestite and barite. 

The metasilicate series, in the regular increase of the 
molecular volumes with the atomic weight, behaves more 
like an isomorphous series, although we know that barium 
metasilicate is not isomorphous with the others. Our 
knowledge of the solid solubility relations of the carbo- 
nate and sulphate series is too incomplete to allow any 
closer comparison. 

In the ref ringence one can hardly see any regular rela- 
tion. The barium compounds have mostly -the highest 
indices of refraction, but the carbonates make an excep- 
tion, aragonite having higher indices than witherite. 
Among the strontium compounds some have higher and 
others lower indices than the corresponding calcium 
compounds, without any apparent regularity. 

Schaef er 37 carried out a thermal study of the binary 
systems CaCl 2 -SrCl 2 and CaCL-BaCL. He found, in the 
former case, a complete solid solubility with a minimum 
in the melting curve, while the latter do not mix at all, but 

36 These values have been obtained by computing the specific gravities at 
25°, compared with water at 25°, as found by Allen and White (this Journal, 
21, 103, 1906), for a -CaO.Si0 2 2.912 and /3-CaO.Si0 2 2.214, in terms of 
density, i. e. comparing with water at 4° and making the correction for the 
buoyancy of air. Thus, for a-CaO.SiOo, d = 2.901 and, for /3-CaO.SiO, 
d = 2.903. 

37 Walter Schaef er, Neues Jahrb. Min. Geol., 1, 15, 1914. 

P. Eskola — Silicates of Strontium and Barium. 371 

form a double compound CaCl 2 .BaCl 2 which shows an 
incongTuent melting, breaking np into crystals of BaCL, 
and liquid. These relations are in all particulars, except 
in the molecular proportion of the double compound, 
strictly analogous to those of a-CaO.Si0 2 to the metasili- 
cates of strontium and barium. 

In the lime-silica system there occur the basic silicates 
3CaO.SiO L , and 3Ca0.2Si0 2 the analogs of which were not 
found in the systems strontia-silica or baryta-silica. The 
last-named system, on the other hand, is the only one in 
which occur silicates more acid than the meta silicate, 
namely, 2Ba0.3Si0 2 and Ba0.2Si0 2 . Among the alkali 
metals, in a similar way, the one having the lowest atomic 
weight (lithium) forms the most basic silicates and those 
with higher atomic weights (potassium, etc.) more acid 


Fig. 17 gives a synoptical view of the melting diagrams 
for the three binary systems CaO-Si0 2 , SrO-Si0 2 , and 
BaO-Si0 2 . To make the diagrams really comparable 
they are all expressed in terms of molecular percentages. 

The compositions and melting or decomposition points 
of the compounds and eutectics, etc. in the three systems 
are listed in table XI. Table XII gives all the impor- 
tant properties determined for the strontium and barium 

In the svstem SrO-SiOo the following compounds were 
found: SrO, 2SrO.Si0 2 ,~SrO.Si0 2 and Si0 2 . With the 
exception of silica each of them was found in one form 
only, although the temperature of formation of the 
strontium silicates was varied from the melting points 
down to about 900°. 

Especial interest was taken in the strontium metasili- 
cate, SrO.Si0 2 , which was found to be closely isomor- 
phous and optically very similar to a-CaO.Si0 2 . It there- 
fore probably belongs to the monoclinic crystal system, 
but its crystals agree so closely with the hexagonal 
system that, judging only from its own properties, it 
would seem to belong to this system. The crystals are 
apparently hemimorphic and might belong to the dihexa- 
gonal pyramidal or, if they are monoclinic, to the mono- 
clinic domatic class. 

372 P. Eskola — Silicates of Strontium and Barium. 

Fig. 17. 

5i0 2 

Fig. 17.— Synopsis of the binary systems BaO-Si0 2 SrO-Si0 2 , and CaO-SiO, 
in mol. %. 

P. Eskola — Silicates of Strontium and Barium. 373 

Table XI. 

Compositions and temperatures in the systems 

CaO— SL, Sr0 2 , and BaO— Si0 2 . 

Tempera- Character of 

Wt. % Mol. % ture °C change 

CaO Si0 2 CaO Si0 2 
Eutectic cristobalite (?)- 

a-CaO.SiO, 37.0 63.0 38.7 61.3 1436 Melting 

a-CaO.Si0 2 48.2 51.8 50 50 1540 Melting 

Eutectic a -CaO.SiO.,-3CaO. 

2SiOo 54.5 45.5 56.3 43.7 1455 Melting 

3Ca0.2SiO, 58.2 41.8 60 40 1475 Decomposition 

Invar, pt. 3Ca0.2SiO,- 

2CaO.SiOo 55.5 44.5 57.3 42.7 1475 Melting 

2CaO.SiO, 65.0 35.0 66.7 33.3 2130 Melting 

3CaO.SiO, 73.6 26.4 75 25 1900 Decomposition 

Eutectic 2CaO.SiO,-CaO.. 67.5 32.5 69.1 30.9 2065 Melting 

CaO 100 — 100 — 2570 Melting 

SrO Si0 2 SrO Si0 2 

Eutectic tridymite-SrO.SiO, 46.5 53.5 33.6 66.4 1358 Melting 

SrO.SiO, 63.2 36.8 50 50 1578 Melting 

Eutectic SrO.SiO,-2SrO. 

SiO, 65.5 34.5 52.5 47.5 1545 Melting 

2SrO.SiO, 77.5 22.5 66.7 33.3 

Eutectic 2SrO.SKX-SrO.. — — 

SrO 100 — 100 — 

BaO Si0 2 BaO Si0 2 
Eutectic tridymite-BaO. 

2SiO, 47 53 25.9 74.1 1374 Melting 

Ba0.2SiOo 56 44 33.3 66.7 1420 Melting 

2Ba0.3Si6 2 62.9 37.1 40 60 1450 Melting 

Eutectic 2Ba0.3Si0 2 -BaO. 

SiO, 65 35 57.8 42.2 1437 Melting 

BaO.S~iO, 71.8 28.2 50 50 1604 Melting 

Eutectic BaO.SiO,-2BaO. 

SiO, 74.5 25.5 53.5 46.5 1551 Melting 

2BaO.SiO, 83.6 16.4 66.7 33.3 

Eutectic 2BaO.SiO,-BaO . . — — — — 

BaO 100 — 100 — 

the compounds BaO, 2BaO. 
Ba0.2SiO„, and SiOo have 

In the system BaO-Si0 2 
Si0 2 , BaO.Si0 2 , 2Ba0.3Si0 2 , 
been found. 

Of these compounds the dibarium trisilicate, 2BaO. 
3Si0 2 , and the barium disilicate, Ba0.2Si0 2 , showed 
very remarkable behavior, being isomorphous, of ortho- 
rhombic symmetry, and forming a complete series of solid 
solutions. The melting diagram of this series belongs 
to Bakhuis Roozeboom's type 1, without maximum or 

374 P. Eskola — Silicates of Strontium and Barium. 





Twinned'on [100] 

Complete series of 
solid solutions with 
a-CaO.Si0 2 

| Complete series 
y of solid 
j solutions. 

Solid solutions 
with anorthite. 

Solid solutions 
with orthoclase. 




o o 

LO ' JO 

t> ; co 

^ • co" 

cm os 

cm * os co co 

j> •• CO OS fc~ 

io* • "^' co co 







I* i i 

H m o 
i — i ft • bo 



good a/5 

perfect /5y 
poor a/5 
and ay 

perfect a/5 
good f3y 
and ay 




1 — 1 


T— I — — 1 
O T-l 

• if* 

fH O 


ft be 

2 V 



o o 



3 O JO ; CD CD 

O ^ CO ^H ! cjd cjd 
OS -2 A CO ^ cS 3 


jo co 

O 00 JO rH 
CO i- ^f CM 

; oo co co co 

T-5 tH t-I t-H 

t-h I co © 

30 ~ 00 O 

CO 3 | JO CO 

t-I II j t-I t-I 



i> CO 
00 *> 


CO * i> CM t-h 0O OS 

t-h ; co o co io jo 

CM • t-I t-I t-I r-< t-I 


i> OS O CO O i> 

« © ^ t— • t— cm os 
t- ©3 ; oo co co jo 

CO co 

^ II 

£- 00 


© T-I 



02 6 


Monocl. ? 

Pseudo (?) 
tabular (0001) 



or needles 


tabular \\a8 


2 5 
S o 
•£ o 





- 1-2 

° %o 

f- 02 t* 
02 CM 02 

9 d 1 9 

02 mh CO CO 

O ° O ° d 

e8 CQ e8 CQ "3 








o 9 


q q 

9 s 

02 CQ 








S rj-J 

-3 CD 


b0 CT 1 

33 3 

rH O 




^O rrj 


o e 

P fl 


•+= -+J 


H3 T3 

Sh SL, 

-bo CJD 

ft ft 

02 02 

P. Eskola — Silicates of Strontium and Barium. 375 

minimum. The indices of refraction show a continuous 
though not linear variation with the composition. 

The strontium and calcium metasilicates form a contin- 
uous series of solid solutions with a minimum in the melt- 
ing curve. The indices of refraction and the densities of 
the mix crystals show continuous variation with the 

The barium metasilicate is not isomorphous with the 
calcium and strontium metasilicates, being optically 
biaxial, with low birefringence and probably of ortho- 
rhombic symmetry. It forms no solid solutions with 
a-CaO.Si0 2 . Instead, there occurs a double compound, 
2CaO.Ba0.3Si0 2 which melts incongruently, breaking up 
into a-CaO.SiO and liquid. This compound is optically 
uniaxial and negative, unlike all the other metasilicates. 
In its refringence it is similar to BaO.Si0 2 . 

Xeither strontium nor barium metasilicate forms, with 
magnesium metasilicate, a double compound analogous to 
diopside, CaO.Mg0.2Si0 2 . This is considered as a 
special case of the more common rule which seems to 
obtain generally in the compounds of the alkaline earth 
elements, namely, that calcium, in those compounds in 
which it can be replaced by magnesia and ferrous oxide, 
cannot be replaced by strontia or baryta, while the iso- 
morphous series including strontium or barium com- 
pounds may have isomorphous and miscible analogs 
among sodium, potassium, or lead compounds. 

Both strontium and barium form feldspars, analogous 
to anorthite. The strontium feldspar has indices of 
refraction exactly like those of anorthite, and the two 
seem to be completely miscible. The barium feldspar is 
monoclinic, like the natural celsian. It forms commonly 
Carlsbad twins of the contact type and the minute twin 
crystals therefore appear as if orthorhombic. 

During this work I had pecuniary assistance, beside 
that from the Geophysical Laboratory, from two funds 
for the advancement of scientific research in my native 
country, Finland, namely, Alfred Kordelin's General 
Trust for the Advancement of Progress and Knowledge, 
and Herman Rosenberg's Traveling Bursaries Trust of 
the University of Helsingfors. 

Geophysical Laboratory, 

Carnegie Institution of Washington, 
Washington, D. C, 
April. 1922. 

376 W. A. Johnston — Sedimentation in 

Art. XXXII. — Sedimentation in Lake Louise, Alberta, 
Canada;* by W. A. Johnston. 


Lake Louise is a small but well-known lake along the 
line of the Canadian Pacific railway in the Canadian 
Rockies. The sediments being formed in the lake are 
glacial silts derived from Victoria glacier near the head 
of the lake. The character of the sediments and condi- 
tions of sedimentation are somewhat similar to those of 
the glacial lakes which existed in northeastern America 
and in other regions at the close of the Ice Age. A 
question of interest therefore arises as to whether sea- 
sonal banding occurs in the sediments and if so, whether 
it is similar to the supposed seasonal banding of the 
glacial lake deposits. Many geologists hold that the 
banding of the glacial lake clays is seasonal in character 
and estimates of post-glacial time have been made by 
counting the annual layers of the post-glacial clays. One 
of the most notable of these estimates is that of Baron 
Gerard de Geer who determined by this method that it is 
about 12,000 years since the ice-sheet disappeared from 
the vicinity of Stockholm, Sweden. 1 There has been, 
however, so far as known, no direct proof of the annual 
character of the layers or bands. In order, if possible, 
to determine this question core samples from the bottom 
of Lake Louise were obtained in June, 1921, by means of 
a bottom sampler and sounding machine, and an estimate 
was made of the average thickness of the annual layer of 
sediment being formed in the lake. 

The character and origin of the lake basin was investi- 
gated and soundings of the lake were made by W. D. 
Wilcox 2 in 1899, and an excellent description of the 
characteristic features of Victoria glacier is given by 
William H. Sherzer in a paper published in the Smith- 
sonian Contributions to Knowledge, vol. 24, 1907. 

* Published by permission of the Director, Geological Survey, Canada. 

1 A geochronology of the last 12,000 years. Extrait du Compte Bendu du 
XI :e Congres Geologique International, 1910. 

2 A Certain Type of Lake Formation in the Canadian Eocky Mountains; 
Jour, of Geology, vol. 7, pp. 247-260, 1899. 

Lake Louise, Alberta, Canada. 377 

The writer is indebted to Mr. Basil Gartom and to other 
officials of the Canadian Pacific railway for information 
regarding Lake Louise and for assistance in the work. 

Victoria glacier. 

Victoria glacier originates at Abbots Pass npon the 
crest of the Great Continental divide and flows nearly 
north between Mounts Lefroy and Victoria to the main 
valley in which Lake Louise is situated, where it turns 
northeast and is joined by the tributary Lefroy glacier. 
The glacier is about 3 miles long measured along either 
branch. It terminates one mile from the lake and at its 
terminus is about 450 feet above the lake. The lower 
part of the glacier, or that part lying in the main valley, 
is covered by morainic material and by slide rock. 
Sherzer showed by observations on steel plates fixed in 
the glacier that the central part moved^in 1904 at a rate 
amounting to 66 feet for the year and by observations at 
the end of the glacier that the forward movement is com- 
pensated for by melting. 3 The glacier is therefore active 
to some extent, — as is also shown by the silt laden waters 
discharged from the glacier, — but is not markedly so. 

Lake Louise. 

Lake Louise is 1% miles long and *4 to % mile wide 
and has a surface area of nearly 1,000,000 sq. yards. It 
has an altitude of 5670 feet above the sea and is 600 feet 
•above Bow River into which it discharges by a small 
stream 2% miles long. The lake is held at its lower end 
by a drift dam, and is bordered on both sides, except near 
its lower end, by rock walls, which rise precipitously for 
considerable heights. The lake has a maximum depth of 
230 feet ; a considerable part of the bottom being nearly 
level. The lake freezes over in the latter part of October 
or early in November and remains frozen until some time 
in June, ice forming to a depth of about 40 inches. 

It is stated by Mr. Gartom that water from the glacier 
flows into the lake even during the coldest part of the 

3 Glaciers of the Canadian Eockies and Selkirks, Smithsonian Contribu- 
tions to Knowledge, vol. 24, p. 32, 1907. 

378 W. A. Johnston — Sedimentation 


winter, when the air temperature at times is as low as 
— 40 degrees F. It does not flow out at the surface but 
seeps through the drift dam at the lower end of the lake 
and issues as springs. 

A delta and alluvial plain is being built at the upper 
end of the lake by the stream flowing from Victoria 
glacier. The delta extends into the lake about 400 feet 
and the alluvial plain extends upstream about 1/3 mile, 
thus indicating the amount of filling of the valley by sedi- 
ment transported by the stream in post-glacial time. 

The temperature of the water in the stream at the exit 
from the glacier is at or slightly above the freezing point. 
At the delta it was 36 to 38 degrees F. in June, 1921 and 
the temperature of the surface water at the lower end of 
the lake was 6 to 8 degrees higher^ At a depth of 6 feet 
below the surface the temperature was 39 degrees. It is 
probable therefore the great mass of the water of the 
lake remains at the temperature of maximum density 
(39.2 degrees F.) through the year. 

Bate of Sedimentation. 

The stream flowing into the lake from the glacier 
carries in suspension the maximum amount of sediment 
at the time of the spring rains, when the snow at the 
lower levels is rapidly melting. It also transports con- 
siderable amounts during the summer but very little dur- 
ing the winter. Sherzer found that the stream at the 
exit from the glacier carried in suspension 506 and 230. 
parts per million of sediment at times of maximum and 
minimum flow in July, 1904. 4 Part of this material, how- 
ever, is deposited on the flood plain of the stream and in 
the part of the delta that is submerged only during high 
water stages. Two samples of the water taken by the 
writer in June, 1921, from the stream at the point where 
it empties into the lake were found to contain an average 
of only 90 parts per million of sediment. The samples 
were taken at a time of approximately minimum flow of 
water for the summer months. As the water in the 
stream contains very little sediment during the winter 
months, the average amount of sediment carried into the 

4 Opus cit. p. 28. 

Lake Louise, Alberta, Canada. 379 

lake by the stream throughout the year probably does not 
exceed 90 parts per million. Most of the sediment trans- 
ported into the lake is deposited on the bed of the lake, 
for the outlet stream is nearly clear. 

The inflow into the lake by the stream from Victoria 
glacier in June, 1921, during a period of moderate rain- 
fall and melting of the ice, was estimated at 85 c. ft. per 
sec. Four determinations by Sherzer, in 1904, upon the 
inflow at the head of the lake gave an average of 80 
c. f. s. ; two at the outlet gave an average of 88 c, f. s., 
the small additional flow coming from Mirror Lake and 
Lake Agnes. 5 The inflow from these lakes is partly by 
underground passages and the water is clear. The 
greater part of the flow from the lake is carried by two 
pipes, one 33 and the other 20 inches in diameter. It is 
stated by Mr. Gartom that the flow during the summer 
months never falls below the capacity of these two pipes. 
During the winter months the inflow is greatly reduced 
as is obvious from the fact that, during the heat of a 
summer day, the flow from the glacier is much greater 
than at night. It is probable, therefore, that the average 
flow throughout the year does not exceed 50 c. f . s. 

Taking the average flow from the glacier as 50 c. f. s., 
the average amount of material carried in suspension by 
the stream as 90 parts per million, the weight of the dry 
silt as 75 lbs. per c. ft. and the area of the lake as 1,000,000 
sq. yds., the average thickness of the annual layer, if 
evenly distributed over the lake bottom, would be nearly 
1 6 inch in thickness. This estimate is of course very 
approximate ; it merely shows that the annual layer, if 
evenly distributed, probably would not exceed 1/6 inch in 
thickness. As the greater part of the sediment is 
deposited near the upper end of the lake, the annual 
layer is thickest near the upper end and thinnest at the 
lower end. The lake is being gradually filled by deposi- 
tion of the glacial silt, but the process is so slow that at 
the present rate of sedimentation it will take over 1000 
years to fill the basin; so that there is no danger of this 
"Gein of the Canadian Bockies," as it is popularly called, 
disappearing in the near future. 

5 Opus cit. p. 28. 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 23.— Xoyember, 19-3'2 

380 W. A. Johnston — Sedimentation in 

Character and Significance of the Stratification of the Sediments. 

Seasonal banding in the sediments being formed in the 
lake is to be expected because of the manner in which 
sedimentation takes place. Considerable quantities of 
silt are carried into the lake during the summer months 
and very little during the winter months. The coarse 
material quickly settles to the bottom, and the very fine 
material remains in suspension, so that by the end of the 
summer the water is slightly turbid and has the greenish 
shade characteristic of silt-laden glacial waters. During 
the winter this very fine material settles to the bottom 
and by spring the water is clear and has the blue color 
characteristic of pure water. 6 There is very little cur- 
rent through the lake, and, therefore, practically all the 
material deposited on the bed of the lake, except in the 
delta, is fine enough to be carried in suspension in nearly 
still water. A small amount of coarse material, which 
accumulates on the ice during the winter, and is derived 
from slides, wind-action or by streams flowing over the 
ice, falls to the bottom when the ice melts. It is to be 
expected, therefore, that the annual layers consist of 
two parts ; that the material is mostly fine-grained and 
that there is a gradation in fineness from the coarser 
lower part of the layer to the upper finer-grained part. 
It is to be expected also, as pointed out above, that the 
thickness of the annual layers in the central part of the 
lake basin does not greatly exceed 1/6 inch. In the sub- 
aqueous part of the delta, at the upper end of the lake, 
conditions are different from those obtaining in the main 
body of the lake. The highly colored, turbid water of 
the creek flows out over the subaqueous part of the delta 
and extends into the lake for varying distances, depend- 
ing upon varying conditions. The water flowing into the 
lake during the summer is nearly at the temperature of 
maximum density, but varies in temperature at different 
times, being colder in the morning than in the evening 
and varying at other times according to weather condi- 
tions. As the great mass of the water in the lake is at 
the temperature of maximum density and as the river 
water usually has nearly the same density — the silt 

6 William H. Sherzer, opus cit. p. 60. 

Lake Louise, Alberta, Canada. 381 

carried in suspension only slightly increasing its dens- 
ity — the river water tends to diffuse through the lake 
water. At times it is lighter than the lake water and 
flows out on the surface over the subaqueous part of the 
delta, but only for a distance of 200 or 300 yards. At 
times also wind-induced currents carry the turbid river 
water out into the lake farther than usual or shift it from 
one side of the delta to the other. There is apparenly 
no tendency, so far as could be observed, for the river 
currents to follow the bottom of the lake and the diffusion 
through the lake of the silt in suspension and the steep 
underwater face of the delta seem to show that no 
appreciable currents follow the lake bottom. 

The material forming the subaqueous part of the delta 
is silt and very fine sand. The very fine sand is probably 
moved along the bottom by current action and comes to 
rest along with the part of the silt which is too coarse to 
remain in suspension, when the current is checked. The 
underwater face of the delta is remarkably steep in its 
upper part, in spite of the fineness of the material. The 
soundings indicate that the slope is in places as much as 
8 degrees. It is probable, therefore, that foreset bedding 
is well developed in the delta and that the material is 
marked by very fine laminations due to short, interrupted 
periods of sedimentation. 

Figure 1 shows two core samples from the bottom of 
the lake, and, by way of comparison, one sample from the 
Champlain glacial clays of the Ottawa valley. The 
material from the bottom of the lake is mostly very fine- 
grained and of a uniform grayish white color. It is de- 
rived mostly from erosion of Cambrian quartzites and 
is therefore highly siliceous in character. Because of the 
unweathered character and uniform color of the samples 
the character of the stratification cannot be well shown 
by the photograph. The stratification is also somewhat 
distorted by the action of the bottom sampler and the 
samples are somewhat compressed but probably not much 
more than the lower compacted beds of the silt. These 
were found to be quite firm at a shallow depth. The 
bottom sampler, weighing 55 pounds, penetrated to a 
depth of only 3 feet or 4 feet when allowed to fall freely 
from the surface. 


W. A. Johnston^— Sedimentation in 

Bottom Sample No. 1 

Lake Louise, Alta. 

Depth 5 Fath. 

Fig. 1. 

Bottom Sample No. 2 

Lake Louise, Alta. 

Depth 35 Fath. 

Section of 

banded silt and clay 

Ottawa, Ont. 

J ; i# 

Fig. 1. — Core samples (Nos. 1 and 2) of sediments being formed in Lake 
Louise, Alberta, Canada, and a section of banded silt and clay from Ottawa, 
Canada. The samples are mounted in plaster of paris and have been sec- 
tioned vertically to show the character of the stratification. 

Lake Louise, Alberta, Canada. 383 

Sample Xo. 1 is from the subaqueous part of the delta 
at the upper end of the lake. In the upper and lower 
portions of the sample exceedingly tine lamination is 
faintly shown. In parts of the sample there is little or 
no trace of lamination. The very fine laminae are ap- 
parently due to short interrupted periods of sedimenta- 
tion — in part probably daily — which are the result of the 
conditions of sedimentation as outlined above. The 
samples from the delta show no definite evidence of 
seasonal banding, except possibly that the parts which 
are finely laminated are summer layers and parts which 
are not laminated are winter layers, but the lack of lami- 
nation may also be due to continuous sedimentation in 
certain parts of the delta during the summer. 

Sample Xo. 2 is from nearly the deepest part of the 
lake at about mid-length. It shows a faint but definite 
banding. The material is all very fine except for a few 
coarse grains of sand which occur occasionally at the base 
of the individual bands. Each band consists of a coarser, 
lighter-colored, lower part which passes upward into a 
finer, darker-colored upper part, the two portions con- 
stituting a band. The bands vary in thickness and aver- 
age between 5 and 6 to the inch. The thickness corre- 
sponds to the approximate estimate of the thickness of 
the annual layer and the character of the banding is that 
which is demanded by the conditions of sedimentation. 
The banding is, therefore, seasonal in character. It is re- 
markably thin but is probably considerably thicker than 
the seasonal layer formed in many of our rock bound 
lakes, whose waters are nearly clear throughout the year. 
It is also thinner than the average banding of many of the 
glacial lake clays of Pleistocene age, but closely resembles 
the banding of the glacial Lake Agassiz clays in the deep 
part of the basin at Winnipeg, Manitoba. 

Sample Xo. 3 is from the base of the Pleistocene marine 
clays in the Ottawa valley and shows a definite banding, 
the darker layers being clay and light colored parts silt 
with some clay and sand. It differs from sample Xo. 2 
in that the clayey layers are more sharply defined and the 
silt layers are much thicker. The material is also 
coarser. The banding is probably seasonal and the 
sharply defined character of the layers may be due to the 

384 W. A. Johnston — Sedimentation in 

deposition in brackish water. The Champlain (late 
Pleistocene) clays in the Ottawa basin are in places well 
banded. The banded clay rarely if ever contains marine 
fossils and was probably deposited in nearly fresh water. 
The clays containing marine fossils, and therefore depos- 
ited in salt water, are not definitely banded. 

Seasonal handing in glacial clays. 

De Geer holds that the banding of the late glacial clays 
is seasonal in character, — the coarser part of the band 
being the summer deposit and the finer part the winter 
deposit, — that the banding is due to the fact that glaci- 
fluvial rivers entering a lake or an only slightly brackish 
inland-sea have followed its bottom, being heavier than its 
water, and that this is clearly shown by the current-bed- 
ding and sand interbedded with the clay ; whereas in the 
case of streams entering the sea, the river water flows 
out on the surface and can only transport the finest clay, 
the coarser sediment being dropped near the shore or ice- 
border, thus giving rise farther out to an almost unlami- 
nated clay. 7 

In the case of Lake Louise, it does not appear, as 
already pointed out, that appreciable currents follow the 
bottom. It is well recognized that the great mass of the 
waters of the glacial lakes of Late Pleistocene time must 
have been at the temperature of maximum density 
through the year, and as silt and clay in suspension only 
slightly affect the density of the water, — unless the silt 
is large in amount and extremely finely divided — it seems 
improbable that appreciable river currents could have 
followed the bottoms of the lakes, except possibly under 
certain conditions. In places where the glacier termin- 
ated in a lake or inland sea and streams under hydrosta- 
tic pressure issued from the glacier, at some depth below 
the surface of the water, it is possible the river currents 
would continue for some distance along the bottom, if 
the density of the river water was slightly greater than 
the lake water. If the water body into which the stream 
flowed was brackish, the river water would be forced to 
rise to the surface because of its less density. The sand 

7 Opus cit. p. 250. 

Lake Louise, Alberta, Canada, 385 

that occasionally forms part of the seasonal layers and 
the current marks may be accounted for in this way as 
suggested by de Geer. Part of the sand, however, may 
be derived from floating ice. Another factor which may 
be of some significance in explaining the current marks 
and minor irregularities in the seasonal bands is that, at 
certain times of the year when the whole water of the 
lake is at the temperature of maximum density, storms 
may disturb the waters down to the bottom and cause 
appreciable bottom currents. 8 Most of the material 
composing the banded clays is fine enough to remain in 
suspension for at least short periods in nearly still water 
and was evidently deposited in quiet water. The lack 
of lamination of the marine clays is probably mainly 
due, as was pointed out in the former paper, 9 to floccula- 
tion in salt water — an electrical phenomenon of surface 
tension — which causes the silt and clay in suspension in 
the river water to settle to the bottom together when the 
two waters are mixed. 

The seasonal layers being formed in Lake Louise are 
much thinner than most of those of the glacial lake 
deposits which the writer has examined. Many of these 
are one-half to one inch or even more in thickness. The 
small thickness of the annual layers being formed in Lake 
Louise may be accounted for by the fact that Victoria 
glacier is only slightly active. The thickness of the 
annual layers would evidently vary according to the dis- 
tances from the source of supply of the material and 
according to climatic conditions. The exact manner in 
which the annual layers were formed probably varied in 
different lake basins. There seems to be little doubt that 
the coarser lower part of the band is the summer deposit 
and the finer upper part the winter deposit, and that the 
difference is due to variations in transporting power of 
the streams in winter and in summer. It has long been 
known that streams issue from some glaciers in winter 
as well as in summer, because of friction raising the 
temperature of the ice above the melting point or because 
of the heat of the earth, but the amounts of material trans- 

s Xeeham and Lloyd : The life of inland -waters, Ithaca, New York, 1916, 
p. 35. 

9 The character of the stratification of the sediments in the Eecent delta 
of Fraser river, British Columbia, Jour, of Geol., vol. 30, p. 128, 1922. 

386 Johnston — Sedimentation in Lake Louise. 

ported by the streams is much greater in summer than in 
winter. Part of the very fine material which forms the 
upper part of the seasonal band may have been carried 
into the lake during the summer and deposited during 
the winter. In other cases, especially in brackish water 
bodies, the finer part was probably transported during 
the winter, for it could not long remain in suspension, if 
the water was even slightly brackish. 

Seasonal deposition in aqueo-glacial sediments has been 
well discussed by E. W. Sayles, 10 who reviews in a well 
illustrated paper the literature on seasonal banding in 
glacial clays, describes the banded clays in the Connecti- 
cut valley and in Rhode Island and discusses the criteria 
for the study of aqueo-glacial sedimentation. He con- 
cludes that the seasonal hypothesis for the banded clays 
is in a very strong position and that there is little danger 
of its being abandoned. The present paper confirms this 

10 Memoirs of the Museum of Comparative Zoology, vol. 47, No. 1, Cam- 
bridge, Mass., 1919. 

Johnston — Imbricated Structure in River-gravels. 387 

Art. XXXIII. — Imbricated Structure in River-gravels; 1 
by W. A. Johnston. 

Gravel deposits formed in river beds and bars fre- 
quently show a characteristic mode of arrangement of the 
coarse gravel and shingle. The stones, for the most part 
dip upstream and overlap or imbricate. This mode of 
arrangement of the stones was figured and referred to by 
James Geikie as imbricated structure. 2 It is only rarely 
described in text books on geology or physiography but 
has probably been noted by many geologists in the field. 
Because it is a criterion which can be used occasionally to 
distinguish marine from non-marine deposits and because 
of the fewness of such criteria, attention is here directed 
to it. Furthermore there seems to be a misconception 
on the part of some geologists as to the character of the 

Imbricated structure was described and its mode of 
formation explained by G. F. Becker in 1892, who stated 3 : 
"If a flattened pebble is dropped into a running stream, 
the water will exert a pressure upon the stone until its 
inertia is overcome, and during this time the pebble will 
tend to swing across the current so as to present its 
greatest area to the pressure. As soon as the resistance 
due to its inertia is overcome, the pebble will sink through 
the water as if the fluid were at rest until its edge touches 
the bottom, and it will then tip downstream until it meets 
support. . . . Many pebbles thus deposited will, with few 
exceptions, be inclined downstream and will rest against 
one another, like overlapping tiles. ' ' This view as to the 
attitude of the stones agrees with Geikie 's description 
and with the present writer's observations. The impor- 
tant points are that the stones tend to dip upstream or, 
as the placer miners sometimes put it, the up-ends of the 
stones are the downstream ends or point downstream, 
and the upstream stones tend to overlap on the down- 
stream stones. In a recent text-book, however, it is 
stated 4 that the pebbles in river beds may imbricate up- 

1 Published by permission of the Director, Geological Survey, Canada. 

2 James Geikie: Structural and Field Geology, 1905, p. 311. 

3 Finite Homogeneous Strain, Flow and Eupture of Rocks, Bull. Geol. Soc. 
Amer., vol. 4, pp. 53-54, 1893. 

4 Lahee, Frederick H., Field Geology, 1916, p. 88. 

388 Johnston — Imbricated Structure in River-gravels. 

stream and dip downstream. This view is apparently 
incorrect. The stones finally assnme a position in which 
they offer the least possible resistance to the current. 
If the stones dipped downstream they would present 
their edges to the current and therefore offer consider- 
able resistance to the current. Moreover field observa- 
tions show that the dip is upstream. It is well recog- 
nized that in hydraulic work it is necessary, in order to 
achieve maximum results, to direct the stream against the 

Fig. 1. 

Fig. 1. — Vertical section of river gravels at Campbellford, Ontario, Can- 
ada, showing imbricated structure. The direction of the current was from 
right to left. 

"stratification" of the stones, that is in a direction oppo- 
site to that of the current which deposited the stones. 
Imbricated structure is frequently used by placer miners 
as a criterion to determine the direction in which the 
streams in abandoned channels flowed and thus to 
attempt to trace the old channels. The inclined attitude 
of the stones is chiefly relied upon for this purpose and 
the further fact that the longer axes of the stones usually 
lie in the direction of the current. Imbricated structure 

Johnston — Imbricated Structure in River-gravels. 389 

is, therefore, of economic importance and its character 
should be clearly understood. 

Becker also pointed out that "On beaches pebbles are 
sometimes imbricated for a few feet in one or another 
direction and sometimes lie nearly flat. The constant 
reversal of the currents due to breaking and retreating 
waves prevents any extensive methodical arrangement, 
and this fact is of assistance in discriminating marine 

Fig. 2. 

Fig. 2. — Vertical section of conglomerate, on Burrard inlet, near Van- 
couver, British Columbia, showing imbricated structure. The direction of 
the current was from left to right. 

gravels from river deposits." 5 This criterion for dis- 
tinguishing marine from non-marine deposits does not 
appear to have received much consideration by geolo- 
gists, probably because of its limited application, but it 
may be occasionally used, as pointed out below, and is of 
importance because of the fewness of such criteria. 

Imbricated structure is usually best shown in the very 
coarse gravel and shingle deposits of river beds, but is 

5 Opus cit. p. 54. 

390 Johnston— Imbricated Structure in River-gravels. 

also shown, in places, in comparatively fine gravel as in 
fig. 1. This figure shows a section of river gravels in 
the bed of Trent river, at Campbellford, Ontario. The 
largest pebbles in the section are about 8 inches in diame- 
ter. The direction of the current was from right to left. 
Imbrication of the pebbles is not well shown because of 
the fineness of much of the material, but the pebbles have 
a fairly uniform dip in an upstream direction and lie with 
their longer axes in the direction of the current. The 
face of the section is vertical except in the lower part 
which is talus. 

Fig. 2 shows a vertical section of Tertiary conglome- 
rate exposed on the south side of Burrard inlet near 
Vancouver, British Columbia. Imbricated structure is 
fairly well shown in the conglomerate and shows that it 
was formed from river gravels. It is therefore non- 
marine. The section is an east-west one, and the stream 
which deposited the gravel flowed west, that is, from left 
to right in the section. The conglomerate beds have a 
general dip of 10 to 15 degrees towards the south. As 
the depositional dip must have been in the direction of 
the current, the dip towards the south is structural. 
Imbricated structure may be occasionally used, therefore, 
both to distinguish marine from non-marine deposits and 
for structural purposes. 

P. Armstrong — Zircon as Criterion, etc. 391 

Aet. XXXI Y. — Zircon as Criterion of Igneous or Sedi- 
mentary Met amorphics; by P. Akmstrong. 

The complete recrystallization of some igneous and 
sedimentary rocks with consequent total loss of their dis- 
tinguishing characters has led petrographers to search 
for some definite criterion by which, in the absence of con- 
clusive field evidence, the genetic origin of such recrystal- 
lized rocks can be accurately determined. Various means 
to this end have been proposed, amongst them the use of 
zircon. 1 This method rests on the assumption that 
zircon, like grains of sand, may undergo rounding 
during water transportation ; those zircons contained in 
igneous rocks always show sharp crystal boundaries. A 
second postulate on which this method rests is that the 
extreme stability of zircon enables this mineral to resist 
physical and chemical changes even under the greatest of 
metamorphic forces. 

Doubts as to the reliability of this method, which has 
been endorsed or employed by various geologists, 2 were 
expressed orally to Professor A. Knopf by several petrog- 
raphers and led to its critical investigation by the writer. 
The conclusions arrived at are briefly set forth below. 

The rocks to be investigated were pounded, not ground, 
in a mortar to pass an eighty-mesh sieve, panned down 
to the heaviest constituents, and dried. Under a binoc- 
ular microscope, equipped with a 24mm. objective, a 
number of zircon grains, up to twenty in some trials, 
were isolated and carefully mounted on glass fibers. 
Each grain was then placed under a No. 4 objective of a 
petrographical microscope and by means of the glass 
fiber, resting on the stage, rotated so as to give a complete 
view of the grain from all sides. In the latter operation 
reflected light was used. 

A preliminary study of four unmetamorphosed sand- 
stones showed that the degree to which zircons are 
rounded diminishes with increasing coarseness of the 

1 J. D. Trueman : The value of certain criteria for the determination of 
the origin of foliated crystalline rocks, Journal of Geology, vol, 20, No. 3, 

2 A. N. Winchell : The Dillon quadrangle, Montana, U. S. G. S., Bull. 574, 
p. 129, 1914. Leith & Mead, Metamorphic Geology, 1915, p. 225 (H. Holt 


Armstrong — Zircon as Criterion of 

sand grains ; some of the coarser sandstones yield zircons 
which under the high power show sharp-edged and lus- 
trous faces. It is in highly zirconiferous sandstones, 

Figs. 1 to 9. 


Figs. 1, 2, 3. — Typical water -worn zircons from zirconiferous sands, Pablo 
Beach, Florida. 

Fig. 4. — Deformed zircon, showing single secondary crystal face; from the 
Stony Creek gneiss (igneous), Connecticut. 

Figs. 5-8. — Zircons from Stony Creek gneiss; all are covered with blistery 
growth and show small secondary faces. 

Fig. 9. — Rounded zircons in a normal, undeformed granite; Norcross 
quarry, Stony Creek, Conn. 

such as those described by Watson and Hess, 3 which are 
evidently the result of repeated re-concentration, that the 
rounding is most highly developed and gives rise to 
shapes characteristic of water-worn grains, such as flat- 

T. L. Watson and F. L. Hess, Univ. of Virginia Phil. Soc, July, 1912. 

Igneous or Sedimentary Met amorphics. 393 

-tened spheres or lenses, kidney or bean shapes, (figs. 1-2), 
and another which by virtue of its general form and 
"eye "-like depressions distributed over its surface may 
be likened to that of a potato (fig. 3). Associated with 
more or less rounded zircons are others in the same rock 
showing smooth, unbroken, and lustrous prism faces. 
Two explanations can be advanced for this fact: That 
the rounded grains had their source farther away from 
their point of sedimentation, or that the prismatic grains 
were transported, enveloped in a protective biotite flake. 
The latter explanation seems to be the correct one, as 
zircon grains attached to biotite have been observed in 
some of the sandstones, by the writer as well as by 
others. 4 

The study of undeformed granites led to the conclusion 
that zircon, the earliest mineral to crystallize out, may 
undergo during consolidation of the rock a certain 
amount of magmatic corrosion, as is evident from a more 
or less pronounced rounding and glassy smoothness, such 
as observed on the edges of a melting cake of ice, of its 
edges, principally those of its pyramidal terminations. 
Furthermore the zircons were found to show a peculiar 
pitting, not in isolated patches, such as might have been 
caused by the process of crushing, but all over the prism 
faces. It was found impossible to distinguish this kind 
of pitting from that seen on sedimentary grains and 
caused by the chipping of the grain during water trans- 
port. That magmatic corrosion of zircon is possible has 
been shown by C. Doelter, 3 who exposed this mineral to 
contact with molten basalt and produced a broad zone of 
corrosion around the zircon grain. The zircons in unde- 
formed granites are generally prismatic, although some 
were found that show a distinctly ovoid form, due to a 
combination of magmatic corrosion and the development 
of numerous vicinal faces. The latter already have been 
described elsewhere. 6 

In examining igneous and sedimentary gneisses the 
fundamental assumption is often made that zircon has 

4 A. Gilligan : The petrography of the Millstone grit of Yorkshire, Quart 
Jour. Geol. Soc, London, vol. 75, No. 300, p. 266, 1920. 

5 C. Doelter, Handbueh der Mineralchemie, Vol. Ill, Part 1, p. 142. 

6 A. Gilligan, op. cit., p. 266. 

394: Armstrong — Zircon as Criterion of 

been absolutely stable under the conditions of tempera- 
ture and pressure to which it was subjected. However, 
the evidence collected during this investigation would 
speak against this assumption. The zircons of igneous 
as well as sedimentary gneisses were found to show a 
peculiar rough surface, blisterlike in appearance, the 
nature of which it was impossible to determine under the 
microscope. In the belief that this blisterlike appearance 
is due to a coating of silica, several grains were treated 
with warm hydrofluoric acid, but without result. The 
biisterlike covering has either destroyed, or, in any case, 
now completely obscures the original crystal face ; how- 
ever, rotation of the grain under the microscope brings in 
view a great number of apparently irregularly placed, 
small, and more or less round faces of rather dull luster 
which appear to have developed on and within the 
blistered surface. Possibly the latter represents a 
decomposition product of zircon and may be one of the 
hydrous forms of this mineral, e.g. malaconite, described 
by Doelter. 7 A zircon, isolated from the igneous Maro- 
mas gneiss of Connecticut consisted of a splinterlike 
plate, extremely irregular in outline but having all its 
edges rounded by the blisterlike covering just mentioned. 
Similarly, parts of prisms were found in other igneous as 
well as sedimentary gneisses, one of their ends showing 
a pyramidal termination, the other a conchoidal fracture, 
the latter modified in exactly the manner as described 
under the Maromas gneiss. It should be of greatest 
interest to determine whether the composition of this 
blisterlike surface is either identical with or related to 
that of the zone of corrosion of Doelter 's, previously 
mentioned. Our conception of the stability of zircon may 
then, perhaps, be greatly modified. 

That zircons are deformed under metamorphic stress 
can be confidently asserted. From an igneous gneiss the 
writer separated a grain, having a curved sausage shape 
and showing on its concave side a single crystal face, 
embedded, so to speak, in the blistery growth which 
covered the rest of the grain (fig. 4). Others show pear 
or club shapes, and many are such perfect spheres that 
they would apparently fully justify the belief that they 

7 Op. eit. 3 p. 136. 

Igneous or Sedimentary Metamorphics. 395 

are of sedimentary origin. (See illustrations, figs. 5-8.) 
All these grains had the peculiar blisterlike covering, 
some, like the sansage-shaped grain, showing also small 
crystal faces, previously described. It is possible that 
the development of the latter is due to local solution and 
recrystallization under stress ; in any case, the blisterlike 
covering seems to have been caused by conditions under 
which zircon was unstable. 

Figure 9 shows three zircons imbedded in, what its thin 
section shows to be, a normal granite of undeformed 
hypicliomorphic granular texture, from the Norcross 
quarry, -Stony Creek, Conn. The roundness of their out- 
lines is striking, and a comparison of these grains with 
those shown by Trueman and Hess in their papers pre- 
viously cited, it is believed, gives convincing proof that 
the origin of a gneiss cannot be determined from the 
study of its zircons in thin sections. 

Since the ultimate source of both igneous and sedi- 
mentary zircons is the same, it was thought that the 
rounding due to abrasion superimposed on magmatically 
corroded zircons in sedimentary gneisses might Jead to a 
means of distinguishing such gneisses from those of 
purely igneous origin. But no such distinction was 
found to be possible. Figs. 5-8 show zircons found in a 
distinctly igneous gneiss. It will be noticed that some of 
the grains are pearshaped and others are more or less 

From the foregoing the conclusion would appear justi- 
fied that the rounding of zircons is no criterion of the 
sedimentary origin of the metamorphosed rocks in which 
they are enclosed, that the degree of rounding due to 
corrasion may not even be large enough, in some of the 
coarser sandstones, to serve as a distinguishing char- 
acter, and that the diagnosis of a rock, so completely 
recrystallized as to obscure the petrological evidence of 
its origin, cannot be effected by the use of zircon as 

Petrological Laboratory, Yale University, 
New Haven, Conn. 

Am. Jour. Sci. — Fifth Series, Vol. IV. No. 23. — November, 1922 

396 C. R. Stauffer—The Minnesota Devonian. 

Aet. XXXV. — The Minnesota Devonian and its Relation- 
ship to the General Devonian Problem of North 
America; by Clinton R. Stauffeb. 

The Devonian of Minnesota has long been known and 
roughly mapped, but a detailed study of the rocks belong- 
ing to this system has been very much neglected. This 
is probably because the area in which it occurs is a drift- 
covered plain that has been only partially dissected by 
erosion and the outcrops that may be found are neither 
frequent nor very satisfactory. But the country has now 
been settled for a longer period than when the early 
surveys of Minnesota counties were made, hence more 
wells have been drilled and more quarries have been 
opened in the region so that the rocks of this system are 
now very much better known than they were thirty or 
forty years ago. Moreover certain parts of the Devonian 
have been found to be filled with fossils thereby making 
the age determination a certainty. 

In all ^bout 1,200 square miles of southern Minnesota 
are covered by Devonian rocks. This area lies in Fillmore, 
Mower and Free Born counties* (1). In the northern 
and western parts of this region, much of the surface is 
comparatively level and well covered by drift so that it is 
not always possible to trace the Devonian border in those 
directions. On the east side it approaches the driftless 
area and the mantle of glacial debris is reduced to a thin 
film, often insufficient to conceal the bed rock, and the 
possibilities for satisfactory stratigraphic work are much 
improved. Enough outcrops can be found to make it 
certain that the Devonian is slightly more extensive in 
this region than it is indicated to be on the present geo- 
logical maps. Numerous masses and fragments of fos- 
siliferous Devonian rock are known or have been picked 
up in the drift of central Minnesota, even as far north as 
Todd and St. Louis counties. Some of these masses are 
quite large (2). This has suggested that possibly there are 
other areas of Devonian, existing as outliers, which have 
not yet been recognized or which may be entirely drift 
covered. It is noticeable, however, that the larger frag- 

* For references see the end of this article. 

C. R. Stauffer — The Minnesota Devonian. 397 

ments are all found in southern Minnesota and may indi- 
cate a somewhat greater extent of the present Devonian- 
covered area rather than the existence of other Devonian 
areas. Over the great ridge area of the buried Minnesota 
mountains the Devonian fragments in the drift are some- 
what smaller and usually less abundant. Hence it seems 
probable that these scattered limestone bowlders and 
loose Devonian fossils have been brought down by the 
Pleistocene glaciers from the great outcrops of Devonian 
in the vicinity of Lake Winnipegosis and Lake Manitoba, 
and that the Minnesota Devonian deposits are confined to 
the southern part of the state. 

The Devonian, as outcropping in the southern part of 
the state of Minnesota, consists chiefly of limestones of 
varying purity. Probably the great body of it runs as 
high as 17% to 18% Mg00 3 but occasionallv layers are 
found with 97% to 98% CaC0 3 and only a fraction of a 
per cent of MgC0 3 . The "best outcrops are to be found 
in the central and southern parts of Fillmore county 
where the Devonian is usually exposed in every highway 
cut. Where both the top and the bottom show, the 
Devonian apparently rests disconformably on fossili- 
ferous Maquoketa shale (Ordovician) and usually has no 
covering other than the drift. In the vicinity of Austin, 
however, the uneven upper surface of the Devonian lime- 
stone is covered by eight to ten feet or more of rather 
soft gray to red clay which has usually been classified as 
a Cretaceous (3) deposit, and which it probably is as clays 
and lignitic beds of that age have been reported in the 
deeper wells of Freeborn county to the west. However, 
some of the similar clays of central and southern Min- 
nesota contain glacial pebbles and are undoubtedly of 
glacial origin. It has been suggested .that a remnant of 
the higher Devonian shales may occur in western Mower 
county and perhaps in certain parts of Freeborn county, 
but up to the present this has not been certainly deter- 
mined. The Western margin of the Devonian is lost 
under a covering of drift which in Freeborn county has 
been estimated to have a thickness of one hundred feet, (4) 
with perhaps even greater thicknesses in the adjacent 
county (5) to the west. There is thus little hope of contin- 
uing the Minnesota Devonian section, except by the drill, 

398 C. R. Stauffer — The Minnesota Devonian. 

to the upper shales which are so well developed a short 
distance to the south of the state line. It seems entirely 
probable that these upper beds thin out in Iowa before 
the Minnesota line is reached. The limestones therefore 
carry the whole of the known Devonian record, as far as 
Minnesota is directly concerned. These rocks dip gently 
to the west and south thus bringing in higher beds along 
a line from northeast to southwest. 

Much of the Devonian is a porous, weathered, impure, 
buff limestone but it changes rapidly in color and 
character as it is followed to the southwest. "Where this 
former character is prevalent it is undoubtedly to be 
assigned, in large measure, to the leaching and weather- 
ing of a rock quite different from that which is now 
exposed. Following the direction of dip there are 
numerous shades and grades between a porous, abund- 
antly f ossilif erous buff rock, and a sparingly f ossilif erous 
blue, or a non-fossiliferous compact gray to white rock. 
A number of sections were measured for detailed study 
and others might have been added. A few of them are 
given herewith to show the Devonian section of the state. 
They include some of the more important outcrops and 
give a good idea of the above mentioned variations as 
well as of the changing character of the fauna which 
seems to attend it. 

Section along the South Bank of Bear Creek at Hamilton , 
Fillmore County, Minnesota. 

Pleistocene and Recent. Thickness 

7. Soil and drift 6' 0" 

Devonian (Cedar Valley limestone) 
6. Limestone, gray to buff, containing the following 

Athyris fultonensis (Swallow) (c) 

Chonetes scitulus Hall (r) 

Productella subalata Hall (a) 

Schizophoria striatula? (Schlotheim) (r) 

Trochonema sp. (r) 15' 4" 

5. Limestone, gray to buff or brown, massive, abund- 
antly fossiliferous. 

C. R. Stauffer—The Minnesota Devonian. 399 

Athyris fultonensis (Swallow) (r) 

Atrypa histryx Hall (c) 

Atrypa spinosa Hall (r) 

Chonetes scitulus Hall (c) 

Cyrtina haniiltonensis Hall (c) 

Gypidula lawiuscula Hall"? (r) 

Martinia sp. (c) 

Productella subalata Hall (a) 

Schizophoria striatula (Schlotheim) (c) 

Spirifer bimesialis Hall (a) 

Spirifer iowaensis Owen? (c) 

Stropheodonta demissa (Conrad) (r) 

Stropheodonta lialli musculosa Cleland (?) (r) 

Bellerophon perplexa AValcott (c) 

Pleurotomaria sp. (r) 

Troclionema sp. (c) 

Coleolus sp. (r) 

Proetns sp. (c) 

....16' 0" 

4. Limestone, partly covered 10' 8" 

3. Covered interval 11' 4" 

Ordovician (Maquoketa shale) 

2. Limestone, buff to brown, shaly, containing Stro- 

phomena fluctuosa Billings 10' 0" 

1. Shale, buff to brown, massive, extending to the 
level of Bear creek. It contains the following 
Dalmanella testudinaria (Dalman) 
Lingula elder i Whitfield 
Plectamonites sericeus (Sowerby) 

10' 8" 

These same Devonian beds ontcrop abundantly around 
Spring Valley. Etna, and southward into Iowa. In 
Spring Valley the old city quarry, at the corner of Church 
Street and Broadway, exposes about eighteen feet of 
slightly higher limestone, but Larsen's quarry in the 
southwestern part of town is essentially the same horizon 
and affords better collecting. 

400 C. R. Stauffer — The Minnesota Devonian. 

Section of Lar sen's Quarry in the Southwestern Fart 
of Spring Valley. 

Pleistocene and Recent. Thickness 

4. Soil and drift 1' 0" 

Devonian (Cedar Valley limestone) 
3. Limestone, gray to brown or buff, badly weathered, 
Atrypa reticularis (Linnaeus) (c) 
Productella subalata (Hall) (c) 

5' 0" 

2. Limestone, buff, massive, with abundant fossils of 
which the following are the more important : 

Atrypa reticularis (Linnaeus) (c) 

Cyrtina hamiltonensis Hall (c) 

Gypidula laeviuscula Hall, (r) 

Productella subalata Hall (aa) 

Nuculites sp. (r) 

Murchisonia cf. dowlingi Whiteaves (r) 

Trochonema sp. (r) 

Trochonema monroei Cleland ? (r) 

Crinoid stems (c) 
1. Limestone, brown to buff, with fossils common 

Chonetes scitulus Hall (c) 

Cranaena iowaensis (Calvin) (c) 

Cyrtina hamiltonensis Hall (t) 

Productella subalata Hall (a) 

Reticularia fimbriata (conrad) (r) 

Spirifer bimesialis Hall (r) 

Spirifer iowaensis Owen? (a) 

Spirifer sp. (r) 

Stropheodonta demissa (Conrad) (c) 

Paracyclas sp. (r) 

Bellerophon sp. (c) 

Trochonema sp. (r) 

Coleolus sp. (c) 

Proetus sp. (c) 

3' 4" 

Somewhat higher beds occur along the state line high- 
way. About five miles w r est of Granger the surface is 
strewn with fragments of the Devonian limestone and in 
some of the fields there are great heaps of rock that have 
been gathered from the surface in clearing the land for 

C. R. Stanffer — The Minnesota Devonian. 401 

cultivation. At no place is this part of the Devonian well 
shown but there is a fair section shown on Mr. Grimm's 
farm where it sticks out along* the road and in the field. 

Section on Mir. Charles Grimm's Farm Five Miles West 

of Granger. 

Recent. Thickness 

4. Soil 0' 6" 

Devonian (Cedar Valley limestone) 

3. Limestone, rough brecciated, brown in color 3' 0" 

2. Limestone, fairly massive, rough, and gray to 
brown in color. It contains occasional masses 
of chert, and cavities filled with calcite crys- 
tals. Fossils are abundant. 

Cladopora magna Hall and Whitfield (r) 

Cladopora sp. (c) 

Favosites sp. (c) 

Zaphrentis solida Hall and Whitfield (a) 

Polyp or a sp. (c) 

Atrypa reticularis (Liunaeus) (c) 

Camarotcechia sp. (r) 

Cyrtina hamiltonensis Hall (c) 

Gypidula laeviuscula Hall? (c) 

Spirifer asper Hall (r) 

Stropheodonta demissa (Conrad) (r) 

Nucleospira sp. (r) 

Schuchertella chemuugensis arctistriata (Hall) (c) 

Conocardium sp. (c) 

Leptodesma sp. (r) 

Palaeoneilo sp. (r) 

Bellerophon perplexa Walcott (a) 

Cyclonema sp. (r) 

Eunema sp. (c) 

Murchisonia dowlingi Whiteaves? (c) 

Pleurotomaria sp. (c) 

Raphistoma terrelli Cleland (r) 

Hyolithes alatus Whiteaves (r) 

Orthoceras sp. (r) 

Poterioceras 2 sp. (r) 

Phacops sp. (r) 

Proetussp. (r) 

10' 0" 

402 C. R. Stauffer — The Minnesota Devonian. 

I. Limestone, gray to brown, rough, with numerous 

corals and stromatoporoids 
Acervularia davidsoni Edwards and Haime (c) 
Favosites sp. (c) 
Stromatoporella erratica (Hall) (c) 

" 3' 0" 

Along 1 the same highway, about two miles farther west, 
fourteen feet of brown to buff limestone outcrop in 
section 34, York Township, Fillmore County, and lie at 
a higher horizon than those beds given in the preceding 
section. Although these layers are not exceptionally 
fossiliferous, the fauna found here is rather large and 
consists of the following forms — 

Idiostroma sp. (r) 
Dictyonema 2 sp. (r) 
Zaphrentis sp. (c) 
Polyp or a 2 sp. (c) 

Semicoscinium rhombicum Ulrich? (r) 
Taeniopora Exigua Nicholson? (c) 
. Atrypa histryx Hall (c) 

Atrypa reticularis (Linnaeus) (c) 
Chonetes scitulus Hall (c) 
Chonetes manitobensis Whiteaves (r) 
Cyrtina hamiltonensis Hall (c) 
Gypidula comis (Owen) (c) 
Pentamerella multicostata Cleland? (r) 
Productella sublata Hall (r) 
Reticularia fimbriata (Conrad) (r) 
Spirifer asper Hall (r) 
Stropheodonta arcuata Hall (r) 
Stropheodonta halli musculosa Cleland? (r) 
Stropheodonta perplana (Conrad) (cc) 
Stropheodonta variabilis Calvin? (r) 
Stropheodonta sp. (r) 
Phacops sp. (c) 

The next beds above this are apparently those out- 
cropping in the vincinity of LeRoy where the following 
section occurs — 

Section of the Fowler and Pay Quarry One Mile East of 
LeRoy, Minnesota. 

Pleistocene and Recent. Thickness 

II. Soil and drift 3' 6" 

C. R. Stauffer — The Minnesota Devonian. 403 

Devonian (Cedar Valley limestone) 

10. Limestone, gray to brown, thin bedded, probably 

somewhat disturbed 4' 0" 

9. Limestone, gray to white, usually weathering to 
brown, fairly massive. It contains a few fos- 
sils among which are the following : 
Crinoid stems (r) 

Leptaena rhomboidalis (Wilckens) ? (r) 
Stropheodonta demissa (Conrad) (r) 
Nucula sp. (r) 
Pleurotomaria sp. (r) 

2' 0" 

8. Limestone, gray to brown, with occasional traces 

of fossils I' 3" 

7. Limestone, gray, fairly fossiliferous 
Crinoid stems (c) 
Athyris fultonensis (Swallow) (c) 
Cyrtina hamiltonensis Hall (c) 
Spirifer orestes Hall and Whitfield (c) 

0' 10" 

6. Limestone, gray to brown, containing indistinct 

masses resembling stromatoporoids 0' 8" 

5. Limestone, gray to white, compact and apparently 

containing no fossils 2' 5" 

4. Shale, green to gray, calcareous, irregular in 

thickness , . 0' 6" 

3. Limestone, gray to white, compact, rather mas- 
sive. It contains a few fragmentary fossils. 2' 4" 

2. Covered interval 1' 0" 

1. Limestone, gray to brown, apparently lower than 

the beds given above » 7' 0" 

Recent drilling in this vicinity indicates that the total 
amount of the high-grade white limestone may exceed 
sixty feet. How much of the brown magnesian limestone 
may be interstratified with it is not definitely known but 
it is probable that there is as much in the deeper strata as 
there is mingled with the beds exposed at the surface. 
Certain layers of the limestone of this locality resemble 
the lithographic beds of Iowa and they may possibly 
represent the same horizon. 

Higher beds occur in the vicinity of Austin. It is quite 
probable, however, that there is a covered interval of 
importance between this outcrop and the one just dis- 

404 C. R. Stauffer — The Minnesota Devonian. 

cussed. There is at least a marked change in sedimenta- 
tion between the light gray to white limestone near LeRoy 
and the impure "cement beds" characteristic of the 
Austin region. The best outcrop in this latter region is 
along Rose Creek, about three miles south of Austin, 
where quarrying has been carried on for many years. 
Formerly this rock was used as a building stone but it 
weathered badly hence it has long been abandoned as a 
construction rock. It is now used only in the manu- 
facture of a natural cement. 

Section of the Fowler and Pay Cement Quarry along 
Rose Creek, Three Miles South of Austin. 

Pleistocene and Recent. Thickness 

4. Soil and drift, the latter chiefly- gravels 2' 0" 

Cretaceous ? 

3. Clay, blue and red, in pockets over the uneven sur- 

face of the limestone 5' 0" 

Devonian (Cedar Valley limestone) 
2. Limestone, blue to gray, weathering to buff, and 
containing a few fossils — 
Athyris fultonensis (Swallow) (r) 
Spirifer iowaensis Owen? (r) 
Fish plates and scales (r) 

10' 0" 

1. Limestone, gray to buff, rather massive, containing 
some chert and with pebble-like masses of 
limestone occurring in some of the lower lay- 
ers. These beds extend to the level of Rose 
creek 15' 0" 

The beds which appear to be the highest of the Devon- 
ian limestone section in Minnesota, outcrop along Cedar 
River in the southwestern part of Mower County. 

Section along Cedar River, Three Miles West of Lyle, 
Mower County. 

Recent. Thickness 

4. Soil 5' 0" 

C. R. Stauffer — The Minnesota Devonian. 405 

Devonian (Cedar Valley limestone) 
3. Limestone, gray to buff, massive to thin bedded, it 
contains an abundance of a few fossil forms — 

Athyris fultonensis (Swallow) (aa) 

Atrypa reticularis (Linnaeus) (r) 

Murchisonia sp. (r) 

4' 0" 

2. Limestone, gray to buff, or brown, rough, hard, 
and massive. Fossils very abundant in some 
layers but limited to one form 

Atrypa reticularis (Linnaeus) (aa) 

10' 6" 

1. Limestone, gray to brown in color, partly covered, 

to the level of Cedar River 6' 0" 

It is quite impossible to present a complete section of 
the Minnesota Devonian at the present time. In fact it 
may never be possible to complete it satisfactorily unless 
at some future time the region should be drilled for water 
or other natural resources that may seem worth while. 
But it may be pieced together, from scattered outcrops 
and other information in somewhat the following manner. 

General Section of the Minnesota Devonian. 
Devonian (Cedar Valley limestone) Thickness 

9. Limestone, buff to brown, massive, coarse, fossili- 

ferous. These are the beds exposed along the 

river just west of Lyle 20' 6" 

8. Limestone, blue to brown, argillaceous, forming the 

cement beds south of Austin 30' 0" 

7. Covered interval 10' ±0" 

6. Limestone, gray to white, compact, fine-grained, 

often alternating with coarser brown beds .. . 22' 0" 

5. Covered interval 12'±0" 

4. Limestone, brown, brecciated, with no fossils 6' 0" 

3. Limestone, gray to buff, massive, f ossilif erous . ... 10' 0" 
2. Limestone, gray to brown, massive, full of corals 

and stromatoporoids 3' 0" 

1. Limestone, buff , massive, abundantly f ossilif erous . 20' 0" 

It is noticeable that all portions of this section are not 
equally fossiliferous and that the fauna is not uniform 

406 C. R. Stauffer — The Minnesota Devonian. 

for the whole formation. Thus the lowest beds are full 
of fossils and the species, the most abundant of which is 
Productella subalata, belong chiefly to the various genera 
of Brachiopoda. These beds are succeeded by others 
which are also quite fossiliferous but the number of 
species is small and they belong mostly to the Anthozoa. 
Then comes the most widely varied fauna of all in the 
beds which have been made the third division of the 
general section. Although the Brachiopoda are still 
abundant the conspicuous forms are Gastropoda and the 
whole fauna differs decidedly from that of both the basal 
beds and those higher in the section. 

The brecciated beds appear to contain few if any 
fossils. The fine-grained compact limestone and the 
associated brown beds are sparingly fossiliferous. The 
argillaceous limestone beds of the cement quarry near 
Austin contain fish remains and an occasional Spirifer, 
while the uppermost portion of the section is again fairly 
fossiliferous but the number of species is small. These 
latter are chiefly Athyris fultonensis and Atrypa reticu- 
laris. The brecciated beds, or the unknown deposits 
which doubtless occur between the outcrops from which 
the section herewith was made, may represent important 
breaks in sedimentation during which marked faunal 
changes occurred, but the evidence so far obtained is not 
sufficient to determine that point. It is perhaps signifi- 
cant that similar faunal changes have been observed in 
the equivalent beds of Iowa (6). 

The following is a list of the genera and species that 
have been collected from the Devonian outcrops of 

Fauna of the Minnesota Devonian. 

Dictyonema 2 sp. 

Idiostroma sp. 

Stromatoporella erratica (Hall and Whitfield). 

Acervularia davidsoni Edwards and Haime. 

Cladopora magna Hall and Whitfield. 

Cladopora sp. 

Favosites sp. 

Hederella filif ormis (Billings). 

Zaphrentis sp. 

Crinoid stem. 

C. B. Stauffer — The Minnesota Devonian. 407 

Polyp or a sp. 

Seinicoscinimn rhonibieum Ilrieli ? 

Taeniopora sp. 

Atkyris coloradoensis Girty. 

Athyris fultoneusis (Swallow). 

Atrypa liistryx Hall. 

Atrypa reticularis (Linnaeus). 

Atrypa spinosa Hall. 

Camarotoecliia sp. 

Chonetes manitobensis AYhiteaves. 

Chonetes seitulus Hall. 

Cranama iowaensis (Calvin) . 

Cyrtina liamiltonensis Hall. 

Gypidula comis (Owen). 

Gypidula lasviuscula Hall?. 

Leptaena rliomboidalis ("Wilckens) ?. 

Martinia sp. 

Xucleospira sp. 

Pentamerella multicosta Cleland?. 

Productella subalata Hall. 

Ketieularia finibriata (Conrad). 

Schizophoria striatula (ScMotlieini). 

Schuehertella enernungensis aretistriata (Hall). 

Spirifer asper Hall. 

Spirifer birnesialis Hall. 

Spirifer euryteines Owen. 

Spirifer iowaensis Owen?. 

Spirifer orestes Hall and TYhitfield. 

Spirifer pinonensis Meek?. 

Spirifer sp. 

Stroplieodonta arcuata Hall. 

Stroplieodonta demissa (Conrad). 

Stroplieodonta halli musculosa Cleland?. 

Stroplieodonta perplana (Conrad). 

Stroplieodonta variabilis Calvin?. 

Stroplieodonta sp. 

Tropidoleptus occidens Hall. 

Conoeardium sp. 

Xucnlites sp. 

Xyassa parva AYalcott. 

Palaeoneilo sp. 

Paracyclas sp. 

Bellerophon perplexa Waleott. 

Belleroplion sp. 

Coleolus sp. 

Cyclonema sp. 

Eunemia sp. 

408 C. JR. Stauffer — The Minnesota Devonian. 

Loxonema sp. 

Murchisonia dovvlingi Whit eaves?. 

Murchisonia sp. 

Pleurotomaria cf . koltubanica Tschernyschew. 

Pleurotomaria sp. 

Raphistoma disciformis Tschernyschew?. 

Raphistoma terrelli Cleland. 

Straparollus clymenoides ? Hall. 

Trochonema monroei Cleland?. 

Trochonema sp. 

Hyolithes alatus Whit eaves?. 

Tentaculites sp. 

Orthoceras sp. 

Poterioceras 2 sp. 

Pliacops sp. 

Proetus sp. 

Fish scales. 

In all the Minnesota Devonian is known to contain sev- 
enty-seven species and some of them are probably new. 
The formation continues southward into Iowa where it is 
said to have a much larger fanna. It is apparently the 
equivalent of the Cedar Valley limestone of Iowa and this 
is a satisfactory formational name in Minnesota because 
much of it is found outcropping in the valley of Cedar 
River and its tributaries. It is possible, however, that the 
limits of the Cedar Valley limestone of Minnesota do not 
correspond exactly to those of the same formation in Iowa 
if the entire Minnesota Devonian is included in that 

The Devonian of Manitoba and the Mackenzie valley 
carry a few of the same fossil forms that are common in 
the Cedar Valley limestone of Minnesota. Unfortunately 
the fauna of that northern region is not as well known as 
might be desired. Whiteaves studies (7) have given a 
very good idea of several divisions especially that of the 
String ocephalus zone in Manitoba and a fair knowledge 
of the same zone at the "Ramparts" on the Mackenzie 
River (8). Although the finding of two drift specimens 
of String ocephalus burtini has been reported in Minne- 
sota (9) the species is not known in the Devonian lime- 
stones of this state. In fact the Cedar Valley limestone 
fauna of Minnesota has nothing in common with the 
String ochalus zone except certain long range species that 
are likely to be found in the Middle Devonian of most any 

C. R. Stauffer — The Minnesota Devonian. 409 

part of North America. A comparison of the Minnesota 
fauna with that of other northwestern Canadian localities, 
as mentioned by Whiteaves (10), gives no more encour- 
aging results. Whiteaves himself says, of his whole list 



Paleogeographic Map of the North American region during late 
Middle Devonian. 

from the Mackenzie valley, ten species are common to the 
Iowa (11) Devonian while twenty-two species are found in 
the Hamilton of Ontario and New York. This is a very 
significant suggestion as to the relationships of the north- 

410 C. R. Stauffer — The Minnesota Devonian. 

western Devonian f annas bnt there are still others equally 
suggestive. Of those common to the Iowa Devonian half 
are general Hamilton forms while most of the others 
belong in the fanna of the Lime Creek shales and are 
hardly to be considered the most characteristic Iowa 
Devonian fossils. And at any rate the Lime Creek fauna 
is quite different from the Cedar Valley fauna and decid- 
edly a later development as far as North America is con- 
cerned. The String ocephalus zone of Manitoba carries 
about 20% Onondaga forms but less than 5% Cedar Val- 
ley species. The later Devonian faunas of the northwest 
are also decidedly different from that of the Cedar Valley 
and appear to have even less in common with it. In 
short, so remote is the relationship between the fauna of 
the Minnesota Devonian and that known from Manitoba 
and the Mackenzie valley that the idea of a direct sea con- 
nection between these two regions, during the deposition 
of the Cedar Valley limestone, should be abandoned. 
Unless the studies at present being pursued in Canada by 
Dr. Kindle, in Iowa by Professor Thomas, and in Mis- 
souri by Dr. Branson, should show a closer relationship 
for the upper beds than is indicated by our present knowl- 
edge it is probable that the supposed sea connection 
across Minnesota during Upper Devonian should also be 

The buried granite ridge, which crosses central Kan- 
sas (12) in a north-northeast direction, was land during 
Devonian time and probably an extension of the land 
area of the Lake Superior region. In fact the pre-Cam- 
brian of this latter region crosses Minnesota as a buried 
ridge and disappears, under the Sioux quartzite, near the 
southwestern corner of the state. While the sea evi- 
dently crossed part of this old mountain range during cer- 
tain periods such as the Upper Cambrian and the late Cre- 
taceous, there are no definite indications that any part of 
it was submerged during the Devonian and the lack of the 
expected relationship between the Devonian faunas on 
either side of the ridge seems to indicate that there was 
a land barrier in that region during the life of these 

The Devonian is represented by 8000 feet of limestone 
and shale in the Great Basin. Its fauna is only partially 
known but "Walcott (13) found it to be a large and varied 

C. R. Stauffer — The Minnesota Devonian. Ill 

one with many similarities to that found typically in the 
Onondaga of New York and Ohio, but Hamilton and 
Chemung species are apparently not lacking in it. This 
relationship cannot be wholly accidental. But perhaps 
the most significant fact about this Devonian deposit is 
that it carries such a large percentage of species not 
known as a part of the eastern fauna. Many of these 
occur in the Iowa Devonian fauna especially in the upper 
beds. But there is still a considerable residue of forms 
most of which have not been specifically identified. It is 
in this latter that hope lies in an attempt to trace the 
Cedar Valley fauna, which probably has its ultimate origin 
in the Devonian of Russia and western Europe, or some 
region which supplied emigrant to all three of these areas. 
About 20% of the Minnesota Devonian species occurs in 
the fauna of the Great Basin Devonian and 30% more of it 
may be the same as those listed by Walcott (14). In the 
6000 feet of limestone, which make up the lower division 
of the Devonian of the Great Basin, nearly the whole of 
Walcott 's collection came from the lower 500 feet, thus 
leaving more than 5000 feet of massive limestone almost 
unexplored and a fruitful field for future research. The 
relationship that exists between the fauna of the Cedar 
Valley limestone of Minnesota and that of the Devon- 
ian limestone of the Great Basin, and somewhat more 
remotely of the middle Devonian of some of the Alaskan 
islands, has suggested the Paleogeographic map which 
accompanies this paper. It is in part a modification of 
one of Professor Schuchert's Devonian maps but it con- 
tains much for which he bears no responsibility. A map 
of this sort can only be suggestive of the conditions as 
they probably existed during any period or epoch. This 
follows from the fact that there has been so much erosion 
during subsequent time and this has probably removed all 
traces of the older deposits over wide areas, while land 
barriers that once existed have been likewise obliterated. 
Such maps are therefore subject to constant revision as 
new facts are discovered and new relationships become 
evident. The most striking fact that comes out during 
this study is the remoteness of the relationship between 
the fauna of the Cedar Valley limestone of Minnesota and 
that of the Devonian of Manitoba. 

Am. Jour. Sci.— Fifth Series, Vol. IV, No. 23.— November, 1922. 


412 C. R. Stauffer — The Minnesota Devonian. 


(1) Geological Survey of Minnesota, vol. I, 1884, pp. 303-307; 357-361. 

(2) Idem, vol. I, 1884, p. 384; vol. II, 1888, p. 184. 

(3) Idem, vol. I, pp. 353-356. 

(4) Idem, vol. I, p. 385. 

(5) Idem, vol. I, p. 460. 

(6) Iowa Geological Survey, vol. XIII, 1903, pp. 268-279. 

(7) Contributions to Canadian Paleontology, vol. I, pt. 4, 1892, pp. 255- 
359, pis. 23-47. 

(8) Idem, pt, 3, 1891, p. 249. 

(9) Eef erred to by Sehuchert in IT. S. G. S., Bull. 87, p. 417. 

(10) Op. cit. pp. 247-250. 

(11) Idem, p. 251. 

(12) Bull. -Am. Assoc. Petroleum Geologists, vol. 4, No. 3, 1920, pp. 255- 

(13) IT. S. G. S., Monograph VIII, pp. 4-8, 1884. 

(14) Idem, pp. 99-211: 274-278; 2S1-285 ; and pis. II-VII; XIII-XVII. 

University of Minnesota, Minneapolis. 

Chemistry and Physics. 413 

I Chemistry and Physics. 

1. A New Method of Separating Arsenic from All Other 
Metals. — L. Moser and J. Ehrlich have been led by theoretical 
considerations to a modification of the well-known method of 
separating arsenic from other metals by the distillation of 
arsenic trichloride. Instead of carrying out this operation in 
a stream of gaseous hydrochloric acid, or also with the vapor 
of methyl alcohol, they employ a stream of air with successive 
additions of strong hydrochloric acid solution. 

The apparatus and the method of operation are very simple. 
A flask of 300 cc. capacny is used as the retort, and this is heated 
by placing it in boiling water up to the neck. A rubber stopper 
with three holes closes the flask, with a glass tube through which 
air is led to the bottom of the flask, with a glass dropping 
funnel supplied with a stop-cock for the introduction of hydro- 
chloric acid solution, and with a bulbed outlet tube which leads 
to the receiver. The latter is a 400 cc. beaker containing 250 cc. 
of water. 

Starting with the substance, with the addition if necessary 
of a reducing agent, such as ferrous sulphate or potassium 
bromide (which has been recommended by Gooch and Phelps 
for this reduction), 50 cc. of hydrochlorine acid (sp. gr. 1.19) 
are added, the apparatus is placed in the hot water, a rapid 
stream of air is passed through, and, after periods of ten minutes 
each, 20 cc. of the hydrochloric acid are added. With 0.15 
to 0.25 g. of AsoO ;3 the whole of it passes over in about 40 
minutes. Very good results were obtained by test-analyses 
where the arsenic was determined volumetrically in the distillate, 
and it was found that no antimony passes over under these 
conditions. — Berichte 55, 437. h. l. w. 

2. .4 New Volumetric Method as Applied to Certain Problems 
in Inorganic Chemistry.— Paul Dutoit and Ed. Grofet have 
devised a method which is unique in furnishing evidence 
of the existence of certain compounds. From a burette which 
is thermally isolated a solution is delivered into a Dewar flask 
containing another solution where a reaction takes place 
between the two dissolved substances. A thermometer graduated 
to 0.01° is read during the titration, and the burette reading 
is plotted against the changes in temperature. Straight-line 
curves are found, with sharp breaks at the ends of reactions. 
Such diagrams show breaks when H 2 S0 4 is half neutralized by 
sodium hydroxide, and also when the normal sulphate is formed. 
The several stages of neutralization of LLPO t by NaOH are 
clearly indicated. Addition of HNO : , to Na 3 P0 4 give curves 

4:14 Scientific Intelligence. 

with breaks at each step of the reaction with the trivalent salt. 
The addition of NaOH to Zn(N0 3 ) 9 gives breaks corresponding 
to the formation of ZnN0 3 OH, of Zn(OH) 2 and of Zn(ONa) 2 
and similar results are obtained with NaOH and lead, mag- 
nesium and copper salts. The salts of Cu, Co and Ni give 
evidence of the formation of successions of complexes as NH 3 
is added. It is to be expeected that this method of investigation 
will be of great value in furnishing a very simple method for 
the investigation of many chemical reactions. — J. chim. pharm., 
19,321 (1922). h. l. w. 

3. Theories of Organic Chemistry ; by Ferdinand Henrich. 
Translated and Enlarged by Treat B. Johnson and Dorothy A. 
Hahn. 8vo, pp. 603. New York, 1922 (John Wiley & Sons, 

The wonderful achievements in connection with the theories 
applied to the carbon compounds are extremely well presented 
in this book, and it furnishes a most excellent source of informa- 
tion for advanced students and teachers of chemistry. The 
translators deserve much praise for making this important 
German work available for the use of English-reading chemists, 
as well as for the introduction of several new chapters and 
other additions which deal particularly with the work of Ameri- 
can investigators, especially with that of the late J. U. Nef 
and of Arthur Michael. The German author has furnished a 
preface to the American Edition in which he approves of these 
additions and modifications. 

The subject is presented historically, with naturally less 
attention to the older, superseded views than to those now pre- 
vailing or under active consideration. The discussions of the 
modern theories are very full and clear, the translation of the 
German text into English appears to be most excellent, and 
the difficult typography involving many, frequently complex, 
structural formulas has been very well done. h. l. w. 

4. The Chemistry of Combustion; by J. Newton Friend. 
12mo, pp. 110. New York, 1922 (D. Van Nostr and Company. 
Price $1.25 net) . — This monograph is the outcome of a series of 
lectures recently delivered by the author in the Birmingham 
Municipal Technical School. It gives a clear and satisfactory 
presentation of the subject in its modern aspects, and since 
it appears that there has been no small text-book of this kind 
to which students may be referred, it may be regarded as 
filling an obvious gap in our literature. 

The first chapter is devoted to definitions, then the phlogiston 
theory is briefly discussed, while the other sections are devoted 
to the combustion of solid carbon, flame, the combustion of 
gaseous hydrocarbons and other gases, ignition temperatures, 
the inflammation of gaseous mixtures and the propagation of 

Chemistry and Physics. 415 

naine in them, and surface combustion. The references to the 
literature are numerous and satisfactory. h. l. w. 

5. Petroleum, Where and How to Find It; by Anthony 
Blum. 12mo, pp. 367. Chicago, 1922 (The Modern Mining 
Books Publishing Company). — This book has been prepared by 
an operator in the "oil business," who has evidently had much 
experience in it, for the benefit of those who are or may become 
interested in this great industry. It is a popular, rather than 
a scientific, book. It presents many interesting facts and 
statistics, and gives much practical information and advice in 
regard to the production of petroleum. h. l. w. 

6. The Heavier Constituents of the Atmosphere. — Sir J. J. 
Thomson has recently applied his method of positive ray 
analysis to several problems involving possibly unknown constit- 
uents of certain gases. For the first case he had been supplied 
by Professor Dewar with a considerable amount of the residues 
obtained by evaporating many thousand tons of liquid air. 
These residues had been absorbed by charcoal and the subsequent 
evaporation of the gases from the charcoal apparently effected 
a fractionation so that only components heavier than krypton 
were retained by the charcoal. 

The positive ray photographs showed the line of xenon very 
prominently and also lines of at least two heavier constituents 
corresponding to atomic weights of approximately 163 and 260. 

The author is of the opinion that these lines are due to mole- 
cules of krypton and xenon, as these numbers are about twice 
the atomic weights of the respective gases. Experiments were 
further made to see if these constituents showed any properties 
analogous to the emanation of radio active substances, but 
evidence of ionization resulting from such presumed emanation 
could not be detected. 

Another application by Professor Thomson of his method was 
to the analysis of the gases obtained from a tube in which 70 mgm. 
of radium chloride had been sealed, after exhaustion of the air, 
for a period of thirteen years. The positive ray line for helium 
was very strong and a faint line corresponding to m/e = 5 
was also found. This latter he ascribes to a compound of 
helium and hydrogen. Neon was not detected. 

A third problem was the testing of gases which had stood over 
radium and also gas lit by deflagrating wires. In these cases 
quadruply charged atoms of nitrogen were detected and triply 
charged atoms of oxygen, nitrogen, and carbon. The compound 
OH 4 invariablv carried a double charge. — Proc. Roy. Soc. 101, 
290. 1922. f. e. b. 

7. The Corrosion of Iron and Steel. — Careful estimates of the 
amount of steel and iron structures or materials which are 
annually rendered unserviceable by rusting place it as high as 

4:16 Scientific Intelligence. 

40 million tons. Whether these figures are exaggerated or not 
the wastage by corrosion is so great that the concerted efforts 
of engineers to produce some form of steel alloy with a capacity 
for resisting corrosion seems imperatively demanded. Apart 
from the known valuable properties of chromium steel which is 
too expensive for use on any considerable scale the most promis- 
ing suggestion has been the introduction of a small percentage 
of copper into a mild steel, which has been thought by several 
investigators to give it a superior resisting power. The report 
of a new series of corrosion tests by Sir Robert Hatfield upon some 
American steels containing from .02 to .27 per cent, of copper 
has just been published. The results of the author's 
observations may be stated as follows: (a) Under atmospheric 
corrosion copper steel was rather less affected than ordinary 
steel especially in the more corrosive condition of an industrial 
atmosphere. The superiority was of the order of 10 per cent, in 
pure air and 25 per cent, under the industrial contamination. 
As is generally the case, material with the rolling scale removed 
was more resistant than with the scale on. (b) In sea water 
ordinary steel corrodes more rapidly at first but the rate of 
corrosion for both materials slows up showing a certain degree 
of self-protective action which was a little greater for ordinary 
steel. The total corrosion of the copper steel however was 
slightly less at the end of 16 weeks than that of ordinary steel. 
(c) In tap water (Sheffield, England) there was little to choose 
between the two materials. Though initially not so corrosive 
as sea water, over a long period it was more corrosive due to 
the absence of any self-protective action in the presence of tap 
water, (d) In a 50 per cent, sulphuric acid bath both materials 
were rapidly attacked at first but whereas the solution of 
ordinary steel continued at a steady rate, that of the copper steel 
showed a very much reduced rate after the scale had been 
removed. The steel containing copper was apparently very 
resistant to a 50 per cent, sulphuric acid solution, (e) A 20 
per cent, sulphuric acid solution showed a more vigorous action 
than the 50 per cent, solution but the superior resistance of the 
steel containing copper was again confirmed. 

The deductions to be drawn from these experiments are (1) 
that the superiority of copper steel under atmospheric corrosion 
is due to and dependent on the amount of sulphurous impurity 
carried by the air; (2) that no advantage will be gained by the 
use of copper steel in ordinary fresh water; and (3) that for 
long immersion in sea water this alloy is. probably inferior. 

The author is further of the opinion that as in the majority 
of service conditions iron or steel is subjected to total or partial 
immersion in natural waters it is by no means certain that a 
copper content as a commercial constituent of mild steel might 
not be deleterious. — Proc. Roy. Soc. 101, 472, 1822. f. e. b. 

Geology. 417 

8. The Mathematical Theory of Probabilities; bv Arne 
Fisher, Vol. I, pp. XXIV, 289. New York, 1922 (The Mac- 
millan Company). — The author who is an actuary by profession 
has written this treatise chiefly for students of statistics, but the 
reader will have to be an expert mathematician to follow the 
analytical development of the theorems. In this second edition 
twelve chapters are devoted to the theory of probabilities of statistics, by which is to be understood such series of 
events as appear in games of chance. Two chapters are given 
to the fitting of various analytical formulas and series to 
statistical data or frequency distributions. The remaining four 
chapters explain and illustrate in detail the method of computing 
the parameters in numerical series. It is difficult to believe 
that any set of data could justify the expenditure of so much 
labor in analyzing the curve into what is after all but an arbi- 
trary set of functions. The author's work is characterized by his 
devotion to the methods of Laplace in the development of the 
theory, and the use of the semi-invariants of Thiele, in preference 
to Pearson's method of moments, in the calculation of the 
parameters of the frequency function. p. E. B. 

II. Geology. 

1. The Paleontology of the Zorritos Formation of the North 
Peruvian Oil Field; by Edmund M. Spieker. Johns Hopkins 
University Studies in Geology, No. 3, 196 pp., 10 pis., 1922. — In 
1867, Mr. E. P. Larkin and Professor F. H. Bradley made 
collections of Miocene fossils in the area of Zorritos, Peru, and 
these were described three years later by Edward T. Nelson. 
Recently far more material was collected in this region by 
Professor Singewald, and all of the known collections are 
here reported on in detail. The Zorritos formation now is 
known to have 44 species of gastropods and 57 of pelecypods. 
Of these, 64 are new. The time appears to be in the main 
Burdigalian, though the higher beds may be of Helvetian age. 
The fauna is a shallow-water one, of warm waters, and cor- 
relates best with similar faunas of Panama and the Antillean 
areas. c. s. 

2. The Recession of the last Ice Sheet in New England; by 
Ernst Axtevs. Amer. Geog. Soc, Research Ser., No. 11, 120 
pp., 6 pis.. 19 text figs., 1922. — In this well printed and edited 
book, the author describes the De Geer method of determining 
the rate of annual deposition of ' ' varved ' ' glacial clays and also 
the rate, of recession of the ice lobes in the lake-filled river 
valleys. This method is, in addition, the only known one for 
measuring earth chronology in actual years. Antevs' work 
relates in the main to the Connecticut valley from Hartford, 

418 Scientific Intelligence. 

Connecticut, north almost to the Canadian border. It took the 
ice something like 4100 years to melt back this distance. How 
long ago the American continental ice sheet began its melting 
away is not yet known. The average recession of the ice appears 
to have been about one mile in 22 years, though locally the rate 
varies from '83 to as much as 1100 feet per year. At times 
there was even a slight re-advance of the ice sheet. 

It took about 5000 years for the ice to melt back from southern- 
most Sweden north for 480 miles. Since the ice melted away, 
another 8500 years has elapsed, so that it is about 13,500 years 
since the continental ice sheet began its recession in southernmost 

The term "varve" is a Swedish word, and in geology 
signifies the annual cycles of sedimentation of glacial clays, 
beginning in the coarser, lighter colored material of summer 
deposition and ending in the darker winter deposit of the finest 
blue muds, having a greasy feel. Varved clays are all laid down 
in fresh-water lakes in front of ice lobes, and when glacially 
derived muds are laid down in the sea, it is said that they are 
not varved but are homogeneous, in that the coarser material is 
mixed with the finest of muds. In fjords, however, such muds 
may also be faintly varved. 

The measuring of the varves and the making of the local 
graphs to show the varying rate of annual deposition is not a 
difficult matter, but requires patience, since the work is both 
laborious and time-consuming. The greatest difficulty of the 
method lies in finding a succession of closely adjacent clay 
(not sand) exposures and in correlating these by means of the 
graphs from place to place. The method tends somewhat to 
underestimate the total time and never to overestimate it. 
Commonly, there are two easily distinguishable layers to a varve, 
but often the lighter colored summer portion will be much 
thicker and more or less banded, while the darker greasy winter 
layer is usually not banded. The normal varve, or the cyclic 
material for one year, is usually much less than one inch in thick- 
ness, but in the vicinity of drainage the summer deposition, when 
Of sand, may rise to as much as 12 feet. 

The De Geer method of evaluating varves also reveals the 
elimatic periodicity of the time of deposition, not only as to the 
short cycles, but the long ones as well. When fossil leaves are 
abundant, one can also discern in years the rate of floral adapta- 
tion and migration. 

The memoir is concluded with a most interesting map by 
Professor Goldthwait, which plots for the area of New York, 
Pennsylvania, New Jersey, and the New England States the 
direction of ice flow, boulder trains, terminal and recessional 
moraines, and the position of the ice-edge for every 100 years, 

Obituary. 419 

as determined by Antevs between Hartford, Connecticut, and 
the Canadian boundary. 

A ready means is now at hand for a definite chronology of 
post-glacial time, and our thanks and congratulations are due 
to Doctor Antevs for his successful results. cs. 

3. A Section of the Paleozoic Formations of the Grand 
Canyon at the Bass Trail ; bv L. F. Noble. U. S. Geol. Survey, 
Prof. Paper 131-B, pp. 23-73, pis. 19-25, 4 text figs., 1922.— In 
this memoir the author brings together in great detail all that 
he has learned about the Paleozoic sequence — Cambrian, Devo- 
nian, Mississippian, Peiinsylvanian, and Permian — of the 
marine and continental strata in the Grand Canyon of the 
Colorado River, during the years 1914, 1916 and 1920. Various 
sections are described from Bass Trail eastward for 35 miles, 
and all of the zones and sections correlated into a generalized 
sequence having a thickness of 4014 feet, besides 506 feet of 
Triassic formations. The Grand Canyon should be the Mecca 
for all stratigraphers, and the worshippers at this grandest of 
Nature's shrines will find guidance and inspiration in Doctor 
Noble's careful study of the sediments deposited here by an 
epeiric sea, from shore to deeper water. c. s. 

4. Essentials for the Microscopical Determination of Rock- 
Forming Minerals and Rocks in Thin Sections: by Albert 
Johaxxsex. Pp. 53, with 24 text figures. Chicago, 111., 1922. 
(The University of Chicago Press, $2.00.) — This work is a 
revision of the author's well-known laboratory manual, "A Key 
for the Determination of Rock-forming Minerals in Thin 
Sections,'** which was published in 1908. The new edition 
appears in a markedly different format, being now in quarto 
instead of octavo, and by rearrangement of the determinative 
tables it has been very notably reduced in bulk — from 542 pages 
to 53. The descriptions of the individual minerals have been 
slightly condensed. Brief notes on the modes of geologic 
occurrence have been added, and the diagnostic differences 
between minerals of somewhat similar optical properties are 
more adequately emphasized. A summary exposition of the 
author's quantitative mineralogical classification of igneous 
rocks has also been added. 

Adolph Kx;opf. 

5. The Rocks of Mount Everest. — The efforts of the members 
of the Mt. Everest expedition of 1922 to reach the Summit of 
the mountain have already been fully given in the public 
press. That it was found possible to reach an altitude of 27,300 
feet, with the aid of oxygen, is sufficiently noteworthy. It is 
still more interesting that a third expedition is already being 
tentatively considered and a greater degree of optimism is 
felt bv the climbers as to ultimate success than after the effort 

420 Scientific Intelligence. 

of 1921 ; this is based practically on the fact that the physio- 
logical effects at altitudes of 26,000 and above were found to be 
less serious than anticipated. 

Dr. A. M. Herron has given the results of the examination of 
rock specimens collected at heights from 23,000 to 27,000 feet. 
The conclusion is reached that "Mount Everest is a pile of 
altered sedimentary rocks — shales and limestones — converted 
into banded hornfels, finely foliated calc-silicate schists, and 
crystalline limestones. The hornfels and fine schists are in the 
field blackish or dark green rocks, conspicuously slabby and with 
a general low dip to the north, which, I believe, adversedly 
and even dangerously affected climbing. The crystalline lime- 
stones are fine-grained pure white rocks. The specimens from 
23,000 and 25,000 feet show in microscope sections a very fine- 
grained aggregate of quartz and a greenish mica, with irregular 
lenticles and veins of chlorite and epidote, and in addition 
sometimes calcite pyrites and sphene. 

"The mountain, from 21,000 to 27,000 feet, is made up of these 
black and dark green rocks, with occasional beds of white lime- 
stone, and veins of quartz and muscovite granite. From 27,000 
to 27,500 feet extends an almost horizontal belt, a sill in fact, 
of schorl muscovite granite, along the whole length of the 
mountain, which rock presumably, by its superior hardness, 
gives rise to the prominent shoulder of the mountain north-east 
of the main peak (shown as 27,390 on Major Wheeler's photo- 
graphic survey map). Above this again are black schists. 
Captain Finch informs me that he saw ammonites at a height 
of about 26,500 feet, but was unable to collect them. 

"As to the age of the rocks forming Mount Everest, they may 
perhaps be assumed, for the present, to be Jurassic or Trias." 
London Geological Journal, September, 1922, pp. 219, 220 ; 
see also July, pp. 67-71, August, pp. 141-144 and October, pp. 
288-^91, with fifteen beautiful reproductions from photographs. 

6. A newly Found Tennessee Meteoric Iron; by G. P. 
Merrill (communicated). — State Geologist William A. Nelson 
has forwarded to the U. S. National Museum a mass of meteoric 
iron recently found by Messrs. CD. McKnight and M. W. Spen- 
cer while working on the roadway leading from Savannah to 
Cerro Gordo, some four miles northeast of the first named town 
in Hardin county, Tennessee. The mass is 18 inches in length, 
roughly dumb-bell shaped and weighs 132 lbs. It is an octahe- 
drite in crystallization and much weathered, undoubtedly repre- 
senting an old fall. A cast will be made of it, after which it will 
be cut and analyzed a portion being retained at the National 
Museum and a portion returned for the State collection at 

7. Minor Faulting in the Cayuga Lake Region; by E. T. 
Long. — The following corrections should be made in the above 
article in the number for April (pp. 229-248). 

Miscellaneous Scientific Intelligence. 421 

The first line on p. 232 should read: The Watkins Glen-Cata- 
tonJc folio deals with, etc. 

Page 233, line 26 from top, Enelinal should read Encrinal. 
Page 236, line 4 from top, north should read south. 
Page 247, line 18 from top, Cayuga should read Cayuta. 
Page 247, line 23 from top, Cayuta should read Cayuga. 

III. Miscellaneous Scientific Intelligence. 

1. Foundations of Biology; bv Lorande Loss "Woodruff. 
Pp. xviii, 476, with '211 illustrations. New York, 1922 (The 
Maeinillan Company). — This text book, designed particularly to 
supplement the laboratory work of college students in the ele- 
mentary course in biology, has been prepared with more than 
usual care both as to scope and proportion. In it the reader will 
find a logical and comprehensive account of the underlying 
principles of the organic world, leading from a simple discussion 
of the physical basis of life through the organization, metabolism, 
reproduction, differentiation, heredity, and adaptation of organ- 
isms, to the evidences of organic evolution. The nineteen chap- 
ters embracing this part of the work are so skilfully correlated as 
to make a continuous and harmonious account of the vital phe- 
nomena in both plants and animals. In the final chapter is told 
the story of the historical development of biological science from 
the earliest times to the present, with a brief account of the work 
of those who have made the most important contributions to the 
subject. A synoptic classification of organisms and a concise 
glossary of technical terms are appended. Many well-drawn 
original diagrams are found among the numerous illustrations. 

The book has so many points of excellence that it is not too 
much to say that its careful reading will give to those who pur- 
sue the subject no further a clear, broad, comprehensive and well- 
balanced conception of life and its evolution, while to those who 
contemplate further work in biology it will furnish an ideal foun- 
dation for their more advanced studies. w. r, c. 

2. The Study of Living Things : A Course in Biology for Sec- 
ondary Schools; by W. H. D. Meter. Boston. Xew York. etc. 
(Ginn and Company). — This laboratory guide is issued in the 
form of a pad of sheets of generous size to be filled out by the 
pupil with answers or drawings according to the special instruc- 
tions on each sheet. Ninety-six exercises, covering the entire 
field of elementary biology, are included, each requiring a practi- 
cal investigation on the part of the pupil and compelling him to 
do some independent thinking. Xo better plan has been devised 
for bringing the pupil into direct contact with the most important 
aspects of the subject. w. r. c. 

422 Scientific Intelligence. 

3. Field Museum of Natural History. Annual Report of the 
Director, D. C. Davies, to the Board of Trustees, for the year* 
1921. Pp. 75, with 16 plates. Chicago, 1922.— This report is 
made of especial interest since it opens with a notice and portrait 
of Dr. Frederick J. V, Skiff, who served as director from Decem- 
ber 16, 1893, until his sudden death on February 21, 1921. What 
Dr. Skiff did towards the development of the museum during a 
period of service extending over nearly thirty years is best appre- 
ciated by those who were closely associated with him. The 
museum has also suffered by the loss of Dr. Frank W. Gunsaulus, 
one of the original trustees, and of Charles B. Cory, curator of 
Zoology. The Museum was reopened in its new building on May 
2, 1921, at that time everything was in readiness of the renewal 
of its work. The entire deficit in the building fund has been 
assumed by President Stanley Field, who had earlier contributed 
the sum of $150,000. Capt. Marshall Field has agreed to contrib- 
ute $50,000 annually for five years to pay for expeditions in the 
field, for new exhibition cases and for the publication of papers 
by members of the staff; he had already contributed $65,000 
toward the deficit in the building fund. 

4. Publications of the British Museum of Natural History. — 
Recent publications are the following : 

Catalogue of the Fossil Bryozoa (Polyzoa) in the department 
of Geology : The Cretaceous Bryozoa, volume IV. This volume 
(pp. 1-404, with 8 plates) by W. D. Lang is part II of the cata- 
logue of the Cribrimorphs, completing the Cretaceous Cribri- 
morph Cheilostomata. See introduction to volume III. 

Catalogue of Books, Manuscripts, Maps and Drawings. Vol. 
VI, Supplement, A to I. Pp. 551, 4to. With Addenda and Cor- 
rigenda to vols. I and II, A to Hooker. Pp. 48. The first vol- 
ume of this Catalogue was published in 1903 ; volumes II- V fol- 
lowed in 1904-1915. These were prepared by B. B. Woodward, 
who, with some clerical aid has. compiled the present Supplement. 

5. National Academy of Sciences. — The autumn meeting of 
the National Academy will be held in New York City, November 
14 to 16. The opening session will be on Tuesday morning in 
Schermerhorn Hall, Columbia University, at 10 o 'clock. The ses- 
sion for the presentation of papers will immediately follow the 
business session. This will be opened by a brief address of wel- 
come from President Butler. So far as possible, papers from 
the Sections of astronomy, chemistry, geology and paleontology 
will be assigned to this day. On Tuesday evening President and 
Mrs. Butler will receive the visiting members and their friends at 
the President's house, 60 Morningside Drive, beginning 8.30. 
A subscription dinner will be held on Wednesday evening, at a 
place to be announced later. 

Geology. 423 

On Wednesday, November 15, the meeting will be held at the 
Rockefeller Institute, Avenue A and 66th Street. So far as pos- 
sible, papers from the Sections of botany, zoology and animal 
morphology, physiology and pathology, anthropology and 
physiology, will be given on this day. The meeting of Thurs- 
day will be held in the auditorium of the United Engineering 
Societies Building. 29 "West 39th Street. Papers from the Sec- 
tions of mathematics, physics, and engineering will be given on 
this day. Thursday evening has been left open for informal 
gatherings of members. The headquarters of the Academy are 
at the Hotel Astor. 11th St. and Broadwav. 


Dr. Alexander Smith, professor of chemistry and head of 
the chemical department in Columbia University from 1911 to 
1921, died in Edinburgh, his birthplace, on September 8 at the 
age of fifty-seven years. Professor Smith held many University 
positions, contributed important researches on the forms of 
sulphur, and (with A. W. C. Menzies) on vapor pressures, wrote 
numerous useful textbooks and in brief was a man of great energy 
and wide influence. 

Dr. F. T. Trouton, emeritus professor of physics in the Uni- 
versity of London, died on September 21 at the age of fifty-eight 
years. Born in Dublin in 1863, he was graduated from Trinity 
College where he early showed his rare keenness of mind. He 
was made at once assistant to the professor of physics, and in 
1902 became Quain professor in University College, London. He 
will be remembered for many important researches, those leading 
to the establishment of Trouton's Law, on Hertzian waves, on 
the viscosity of solids and others of no less importance. 

Dr. David Sharp, the veteran English entomologist, died on 
August 27 at his home in Brockenhurst, at the age of eighty-two 

Dr. William Kellner, the eminent chemist, died on Septem- 
ber 12, in his eighty-third year. He was born and received his 
education in Germany, but came to England in 1862 as assistant 
to Sir Henry Roscoe at Manchester. He was made chemist to the 
British War Department in 1902. 

Dr. Arthur Lalanne Kimball, for thirty-one years professor 
of physics at Amherst College, died on October 22 at the age of 
sixty-six years. 

Dr. Albert Averx Sturley, instructor in physics in Yale Uni- 
versity, died on October 22 at the age of thirty-five years. 


Ward's Natural Science Establishment 

A Supply-House for Scientific Material. 

Founded 1862. Incorporated 1890. 

A few of our recent circulars in the various 
departments : 

Geology: J-32. Descriptive Catalogue of a Petrographic Col- 
lection of American Eocks. J- 188 and supplement. 
Price-List of Rocks. 

Mineralogy: J-220. Collections. J-238. Minerals by Weight. 
J-224. Autumnal Announcements. 

Paleontology: J-201. Evolution of the Horse. J-199. Palae- 
ozoic index fossils. J-115. Collections of Fossils. 

Entomology: J-33. Supplies. J-229. Life Histories. J-230. 
Live Pupa?. 

Zoology: J-223. Material for dissection. J-207. Dissections 
of Typical Animals, etc. J-38. Models. 

Microscope Slides: J-189. Slides of Parasites. J-29. Cata- 
logue of Slides. 

Taxidermy: J-22. North American Birdskins. Z-31. General 

Human Anatomy: J-37. Skeletons & Models. 

General : J-228. List of Circulars & Catalogues. 

Ward's Natural Science Establishment 

84-102 College Ave., Rochester, N. Y., U. S. A. 


number consisting of 100 to 120 pages). Editor : EUQENIO RIGNANO. 

This is the only review which has a really international collaboration ; which is 
of world-wide circulation and occupies itself with the synthesis and unification of 
knowledge, in the history of the sciences, mathematics, astronomy, geology, physics, 
chemistry, biology, psychology and sociology. 

It Is the only review which, by inquiries among eminent scientists and writers 
(on: The philosophical principles of the sciences; Fundamental astronomical and physical 
questions of current interest; Contributions given by the various countries to different 
branches of knowledge; Question of vitalism; Social question; International questions 
raised by the world war), makes a study of the problems interesting scientific and intel- 
lectual circles throughout the world. 

It has published articles by Messrs. : Abbot - Arrhenius -Ashley - Bayliss - Beichman- 
Bigourdan - Boh I in - Boh n = Bonne sen - Borel - Booty - Bragg - Bruni - Burdick - Carver - 
Caullery - Chamberlin - Charlier = Claparede - Clark - Costantin - Cromtnelin = Crowter - 
Darwin - Delage - De Vrles - Durkheim - Eddington - Edgeworth - Emery •= Enrlques - 
Fabry - Findlay - Fisher - Fowler - Golgi - Gregory - Harper- Hartog -Heiberg - Minks - 
Hopklns-Inlgaes-Innes=Janet«Kaptein-Kaye-Kidd-Langevin=Lebedew=Lloyd Morgan - 
Lodge - Loisy- Lorentz =■ Loria -Lowell - MacBride - Meillet =lYloret-M uir=Peano -Picard - 
Polncare - Palseux - Rabaud - Rey Pastor - Righi = Rignano=Russell-Ru therford=Sagnac - 
Sarton- Schiaparelli - Scott - See - Sherrington - Soddy - Starling- Svedberg- Thomson - 
Thorndike-Turner -Volterra -Webb -Weiss -Zeeman and more than a hundred others. 

" SCIENTIA" publishes its articles in the language of its authors, and joins to the 
principal text a supplement containing the French translations of all the articles that 
are not in French. ( Write for a Specimen Number to the General Secretary of 
"Scientia", Milan, sending, -to defray postal and other expenses,- 2 Francs in 
stamps of your country.) Annual subscription : 40 sh., or 10 dollars post free. 

Office : 43 Foro Bonaparte, Milan, Italy 

Publishers: WILLIAMS & NORG ATE - London ; FELIX ALCAN- Paris 


Art. XXXI. — The Silicates of Strontium and Barium: by 

P. Eskola, , 331 

Art. XXXII. — Sedimentation in Lake Louise, Alberta, 

Canada; by W. A. Johnston, . , 376 

Art. XXXIII. — Imbricated Structure in River-gravels; by 

W. A. Johnston, -. 387 

Art. XXXIV. — Zircon as Criterion of Igneous or Sedimen- 
tary Metamorphics ; by P. Armstrong, 391 

Art. XXXY. — The Minnesota Devonian and its Relationship 
to the General Devonian Problem of North America; by 
C. R. Stauffer, 396 


Chemistry and Physics. — A New Method of Separating Arsenic from All Other 
Metals, L. Moser and J. Ehrlich: A New Volumetric Method as Applied 
to Certain Problems in Inorganic Chemistry, P. Dutoit and E. Orofet, 
413. — Theories of Organic Chemistry, F. Henrich: The Chemistry of 
Combustion, J. N. Friend. 414 — Petroleum, Where and How to Find it, 
A. Blum: The Heavier Constituents of the Atmosphere, J. J. Thomson: 
The Corrosion of Iron and Steel, 415. — The Mathematical Theory of Prob- 
abilities, A. Fisher, 417. 

Geology. — The Paleontology of the Zorritos Formation of the North Peruvian 
Oil Field, E. M. Spieker: The Eecession of the last Ice Sheet in New 
England, E. Antevs, 417. — A Section of the Paleozoic Formations of the 
Grand Canyon at the Bass Trail, L. F. Noble: Essentials for the Micro- 
scopical Determination of Minerals . and Rocks in Thin Sections, A. 
Johannsen: The Rocks of Mount Everest, 419. — A Newly Found Tennes- 
see Meteoric Iron, G. P. Merrill: Minor Faulting in the Cayuga Lake 
Region, E. T. Long, Errata, 420. 

Miscellaneous Scientific Intelligence. — Foundations of Biology, L. L. Woodruff: 
The Study of Living Things: A Course in Biology for Secondary Schools. 
W. H. D. Meier, 421. — Field Museum of Natural History, Annual Re- 
port for 1921: National Academy of Sciences, 422. 

Obituary — A. Smith: F. T. Trouton: D. Sharp: W. Kellner: A. L. Kimball: 
A. A. Sturley, 423. 

Library, U. S. Nat. Museum. 


Established by BENJAMIN SILLIMAN in 1818. 



Editor: EDWARD S. DANA. 



op Cambridge, 



FREDERICK E. BEACH, of New Haven, 

Professor EDWARD W. BERRY, of Baltimore, 

of Washington. 



No. 24— DECEMBER, 1922. 


19 22. 


Published monthly. Six dollars per year, in advance. $6.40 to countries in the 
Postal Union ; $6.25 to Canada. Single numbers 50 cents; No. 271, one dollar. 

Entered as second-class matter at the Post Office at New Haven, Conn., under the Act 
of March 3, 1879. 

Wiley Books 

for Mineralogists and Geologists 

Published During 1922 


Elements of Optical Mineralogy— Part I, Principles and 

Second Edition, Entirely Rewritten and Much Enlarged. 
By ALEXANDER N. WincheLL, Doct. Univ. Paris. 
An introduction to microscopic petrography, adapted to the study of 
minerals, both in thin sections and in powdered form immersed in liquids. 
216 pages. 6 by 9. 215 figures. Cloth, $2.50. 

Coal : Its Properties, Analysis, Classification, Geology, Extraction, Uses 
and Distribution. 

By ELWOOD S. Moore. 

Anyone desiring to obtain a general knowledge of the subject of Coal can 
do so with the aid of this book. 

462 pages. 6 by 9. 142 figures. Cloth, $5.00. 

to date. 

A Text-book of Mineralogy, By Edward S. Dana. 
Third Edition, Revised and Enlarged, By William E. Ford. 

The changes made consist chiefly of additions bringing the work right up 

720 pages. 6 by 9. 1050 figures. Cloth, $5.00. 

Mineral Land Surveying. Third Edition. 
By James Underhill, Ph.D. 

This new edition has been revised to represent present-day practice in the 
western portion of the United States. 

237 pages. 4£ by 7. 43 figures. Flexible, $3.50. 

Send for any Wiley book on Free Examination ter: 


432 Fourth Avenue 
New York 

London Montreal, Quebec 

CHAPMAN & HALL, Ltd. Renouf Publishing Company 

AJS 12.22 




Art. XXXVI. — John Day Felidce in the Marsh Collec- 
tion;, by George F. Eatox. 

[Contributions from the Othniel Charles Marsh Publication Fund, Peabody 
Museum, Yale University, New Haven, Conn.] 

To the surviving personal friends of 0. C. Marsh, and 
also to the younger generation of vertebrate paleontolo- 
gists, Marsh's eagerness to secure fossils from newly 
reported localities in the western United States, and the 
success of the collectors he employed, no longer present 
any novelty. It need therefore occasion no surprise 
that the Marsh Collection of Vertebrate Fossils in the 
Peabody Museum of Yale University should contain a 
considerable amount of material, from the John Day 
Valley of Oregon, that was collected principally between 
the years 1870 and 1877. The present short article is 
based on that portion of this material from the John Day 
Valley which includes the Felidae. 

Virile it would be ungracious to criticise the methods 
of collecting in vogue in the seventies of the last century, 
it appears that the superior value of a good skull over the 
other skeletal parts, which were then somewhat dis- 
paragingly termed "joints," was a little over-emphasized 
in Marsh's instructions to his collectors. A result of 
this is possibly to be seen in the predominance of cranial 
material, unaccompanied by other skeletal parts that 
might otherwise have been saved, and that would now 
greatly enhance the value of the collection. Then, too, the 
stratigraphy of the region was neither so well under- 
stood nor regarded as of the same importance in con- 
nection with the search for fossils as it has been of recent 
years, and accordingly it would be unwise to accept as 
authoritative, in some instances, the recorded statements 
of the collectors regarding the horizons from which speci- 
mens were obtained. For this reason the present article 

Am. Jour. Sci. — Fifth Series, Vol. IV, No. 24.— December, 1922. 

426 Eaton — John Day Felidce in Marsh Collection. 

must be limited, in the main, to a discussion of the 
osteology and affinities of the material examined, and 
less attention can be paid to stratigraphic considerations 
than would otherwise be desirable. Yet notwithstanding 
real or suspected shortcomings in field work during 
pioneer days, the fact remains that this material, collected 
half a century ago, is worthy of careful study, and con- 
tains forms specifically distinct from those previously 
described. It is a pleasant duty therefore to express 
the admiration that is felt for the generous enthusiasm 
that led Professor Marsh to be one of the first paleontolo- 
gists to visit the John Day Valley and to encourage the 
search for fossils in that region. 

Tables. — For convenience in comparing the new 
species, here described, with the earlier types, cranial 
measurements have been arranged in parallel columns in 
Table A (page 442). While most of the measurements, 
selected for this purpose, conform with those used by 
previous authors, three new measurements have been 
introduced, and an external palatal length has been 
substituted for the old mid-line measurement, This 
table as originally constituted included practically all 
the measurements of, cranium and mandible used by 
Merriam and by Thorpe ; but while extremely useful in 
the study of material in the laboratory, it was thought 
too large for convenience of publication at the present 
time. The table has therefore been cut down to about 
one third of its original size by the omission of all mandi- 
bular and nearly all dental measurements, as well as of 
several cranial measurements that, on trial, did not yield 
satisfactory differential indices. Although the number 
of measurements and indices has been reduced, it has 
seemed desirable that the series of specific types, and 
other crania, tabulated under the several genera, should 
be as complete as possible, in order that the range of the 
indices characterizing these genera might be reliably 
determined. For this purpose four of the best preserved 
crania of the genus Nimravus in the Marsh Collection 
have been measured and the resulting indices recorded. 
Examination of several of Cope's types, now in the 
American Museum of Natural History in New York, 
has been possible through the courtesy of Dr. W. D. 
Matthew of that institution: and Professor John C. 

Eaton — John Day Felidcc in Marsh Collection. ±*27 

Merriam, lately of the University of California, has very 
kindly supplied certain measurements of his types of 
Pogonodon davisi and Nimravus {Archodurus) debilis 
major in addition to those given in his original 
description of these two types. The illustrations accom- 
panying this article are by Mr. R. Weber, and have a 
value of their own quite apart from the text. 

Where the number of species to be compared is large, 
interpretation of the absolute measurements is facilitated 
by a further table expressing relative cranial pro- 
portions. The system of indices, so extensively used 
by anthropologists, eliminates the factor of mere size and 
is admirably adapted to this requirement. Accordingly 
the basal length, measured from basion to prosthion, 
which is probably the most rational single element of 
size, has been chosen as a convenient basis for the com- 
parison of the major cranial proportions. From several 
of the cranial measurements in Table A, indices have 
been derived by the simple process of dividing these 
measurements, as taken on each skull, by the basal length 
of that skull, and multiplying the quotient by 100. In 
the case of minor cranial proportions and dental pro- 
portions, other dimensions than the basal length are more 
suitable as the bases of comparison ; but as in every case 
the definitions of the indices, arranged in Table B, refer 
to the serial numbers of the absolute measurements, 
recorded in Table A, no confusion regarding the identity 
of the indices in Table B need arise. 

The measurements recorded in Table A call for little 
explanation. The axial length from the prosthion to the 
posterior surfaces of the occipital condyles has been 
retained principally because of its long use by previous 
authors as a convenient over-all dimension ; but the basal 
length from basion to prosthion, which can be readily 
taken in some skulls where Cope's axial length is impracti- 
cable, has been preferred as a basis for the comparison of 
size and proportion. There has been little uniformity in 
the past regarding the measurement of palatal length. 
The margin of the palatine bones bounding the posterior 
nares is seldom perfectly preserved, and even when com- 
plete it seems to be subject to some individual variation of 
form due to the development of a posteriorly directed 
mid-line process homologous with the posterior nasal 

428 Eaton — John Bay FelidcE in Marsh Collection. 

spine of the human skull. It has seemed desirable 
therefore to discard the mid-line palatine length, and to 
substitute for it a palatal length measured from the 
prosthion to a line tangent to the posterior surfaces of 
the maxillary parapets. This is precisely the same as 
the external palatal length of human craniometry. 
Since the index of palatal length, computed from this 
dimension and the basal length, is remarkably constant 
in all the species of Nimravus, Binictis, and Pogonodon 
examined during the preparation of these tables, the 
palatal length furnishes an excellent basis for computing 
indices of palatal breadth. 

Pogonodon Cope. 

This genus was founded by Cope 1 on the characters 
of a skull that he had originally described as Hoplo- 
phoneus platycopis. 2 In a later article 3 he stated that 
"This genus [Pogonodon'] represents a station on the line 
connecting Binictis with the higher sabre-tooths, being 
intermediate between the former genus and Hoplo- 
phoneus," and provisionally referred to the same genus 
incomplete material which he had made the type (partim) 
of Machcerodus brachyops. 4 " Pogonodon platycopis and 
P. brachyops were transferred by Adams 5 to Binictis, 
the reason for this change being that "The genus Pogono- 
don as proposed by Cope does not differ from Binictis 
as regards tooth structure, and the absence of the second 
inferior molar, which in Binictis is much reduced, is not 
sufficient grounds for retaining it as a distinct genus, 
since in several specimens of Binictis it is variable in 
size and in one is absent from one side." Matthew 6 
referred Pogonodon platycopis to Cope's genus Nimra- 
vus, regarding Pogonodon as a distinct sub-genus ; and 
to the same genus he referred P. brachyops also, stating 
his opinion that the latter species was "like N[imravus] 
gomphodus and doubtfully separable from it except by 
absence of the infracarnassial exostosis." In the same 

1 E. D. Cope, Am. Nat., 14, 143, 1880. 

2 E. D. Cope, Am. Nat., 13, 798a, 1879. 

3 E. D. Cope, Am. Nat., 14, 847, 1880. 

4 E. D. Cope, Proc. Am. Phil. Soc, 18, 72, 1878. 
5 C. I. Adams, This Journal (4), 1, 433, 1896. 

6 W. D. Matthew, Bull. Am. Mus. Nat. Hist., vol. 28, 310, 1910. 

Eaton — John Day Felidce in Marsh Collection. 429 

article, in his discussion of Hoplophoneus, Matthew notes 
that Merriam 's Pogonodon davisi, which Matthew had 
not then had an opportunity to examine, "appears from 
Merriam's figures to be referable to this genus [Hoplo- 
phoneus] and distinct from any of the Dinictis phylum.' ' 
On the other hand, Merriam 7 in his announcement of 
Pogonodon davisi has stated at considerable length the 
reasons for which it seemed to him "advisable to use the 
arrangement proposed by Cope, and to separate platy- 
copis, bracliyops, and davisi [from the deinictid group 
in the John Day beds] as the Pogonodon group, of at 
least subgeneric rank." In this status Pogonodon has 
rested for several years. My colleague M. R. Thorpe 8 
has recently found it expedient to recognize the validity 
of Cope's Pogonodon in order satisfactorily to record 
the affinity of a feline mandible from the White River 
beds of South Dakota. Similarly in the case of the skull 
about to be described, the practical convenience of 
retaining Pogonodon as a distinct genus, or subgenus, 
is not to be denied. 

Pogonodon serrulidens, sp. nov. 

(Figs. 1-3.) 
Holotype, Cat. No. 10520, Y. P. M. Upper Oligocene (John Day), Turtle 
Cove, John Day Valley, Oregon. Collected by L. S. Davis in 1875. 

The type material consists of a cranium, not particu- 
larly well preserved, together with two fragments of 
the right mandibular ramus, the proximal portion of the 
right ulna, and the proximal portion of the left metatarsal 
III. From the indications of age afforded by the sutures 
alone, this skull would appear to be that of an animal 
nearly or quite adult, but the dentition proves that the 
animal was immature. The permanent premolar and 
molar series is, or was, complete, while the deciduous 
canine of the left side is still present, being but partly 
overlapped, on its inner side, by its permanent successor. 
On the right the deciduous canine has been lost with 
part of the maxilla, so exposing considerably more of 
the permanent canine than the portion that had actually 
protruded beyond the parapet. The premaxillary 
alveolar border has been damaged, and none of the 

7 J. C. Merriam, Univ. of Calif., Bull. Dept. Geol., vol. 5, 57, 1906. 
8 M. E. Thorpe, This Journal (4), 50, 223, 1920. 

430 Eaton — John Day F elided in Marsh Collection. 

incisors are preserved. While the immature develop- 
ment of the skull admittedly detracts somewhat from its 
value as a specific type, this deficiency is nearly compen- 
sated by the fact that the teeth of the permanent set 
exhibit practically no signs of wear, and therefore 
present, so much the more clearly, their characteristic 

The material is provisionally attributed to the upper 
John Day because of the gray color of the ashy matrix, 
which, although slightly tinged with green, can not be 
described as bluish-green; but as this stratigraphic 
reference is unsupported by field records or other data, 
it should not be regarded as established beyond question. 

Fig. 1. 

Fig. 1, — Pogonodon serrulidens, sp. n<Xv. Holotype. X %• 

Dentition. — I 3 , C 1 , P 3 , M 1 . The upper canines are 
compressed. The anterior margin of the right upper 
canine, which is exposed to view for nearly half its 
length, is sharper than in Nimravus, the serrated anterior 
ridge following the actual anterior margin of the tooth 
for a greater distance from the point than is the case in 
Nimravus. P 2 is very small, practically vestigial, and, 
single-rooted. P 3 is much smaller in proportion to P 4 
than in Dinictis and Nimravus, in this respect nearly 
resembling Pogonodon davisi. P 4 is without a positive 
inner cusp (protocone), the inner root being small and 
closely adpressed to the anterior root. It has a small 
but definite anterior accessory cusp (parastyle) distinctly 
separated from the sharp anterior margin of the para- 

Eaton — John Bay F el idee in Marsh Collection. 131 

cone by a narrow oblique cleft. This is a character which 
the present species shares with Pogonodon davisi, but not 
with P. plat y co pis, for according to Cope 9 the upper 
sectorial of the latter species has no parastyle. The 
paracone and metacone of P 4 are separated by a deep 
notch much as in Dinictis and Nimravus j and judging 
from the illustrations of Pogonodon davisi, as in that 
species also. M 1 is small with much reduced inner lobe. 
P 4 which is fortunately preserved in one of the mandib- 
ular fragments, has a relatively greater transverse 
breadth than I have observed in specimens of Nimravus. 
In this greater breadth and in the extent and form of the 
notch separating the posterior basal tubercle from the 
principal cusp, this tooth closely resembles the P 4 of 
Thorpe's Pogonodon cismontanus, Cat. No. 10053, 
Y. P. M. 

The cutting edges of all the teeth, including even the 
vestigial P 2 , are finely serrated. This condition would 
facilitate the sectorial action of the teeth of a young 
and comparatively weak animal. It would be of 
progressively less advantage as the jaws and their mus- 
cles acquired greater strength. By the time the animal 
was fully adult, the serrations would probably be nearly 
or quite worn away from all the cheek-teeth, while on the 
canines, which because of their special function are less 
subject to detrition, the serrations might be expected 
to persist much longer. Sand and gritty substances 
taken into the mouth with food are a constant cause of 
dental detrition among the Carnivora, not excepting even 
the cleanly Procyon lotor, and in view of the habit of some 
recent large felines of crunching bones, 10 it is probable 
that similar habits, acquired as the large extinct felines 
approached maturity, would hasten the destruction of 
the serrated sectorial edges. The resemblance of these 
serrated teeth to small saws has suggested the specific 
name, but it is not assumed that the cheek-teeth were 
serrated only in this species, or indeed in this genus alone. 

One of the mandibular fragments supports a lower 
permanent canine lacking the upper half of the crown. 

9 E. D. Cope, Eeport of the TJ. S.. Geol. Survev of the Territories, vol. 3, 
981, 1884. 

10 Dr. W. Eeid -Blair, D. V. S., of the New York Zoological Park, informs 
me that the lion, tiger, leopard and prima all use the large cheek-teeth for 
this purpose. 

432 Eaton — John Day Felidce in Marsh Collection. 

The tooth is not quite fully protruded. The order of 
dental replacement, exhibited in this skull, should be 
noted. Disregarding the incisors, none of which have 
been preserved, there are actually in place and functional, 
in the lower jaw, a permanent canine and a permanent 
P 4 , and in the. upper jaw, on one side, all the permanent 
premolars and molars. On the left side the alveolar 
margin where P 2 should be has been destroyed. From 

Fig. 2. 

Fig. 2. — Pogonodon serrulidens, sp. nov. Holotype. X %• 

the foregoing it appears that the deciduous upper canines 
remained in place until all the permanent upper cheek 
teeth were fully erupted. The presence of P 4 in the 
mandible renders it probable that the replacement of 
the lower cheek-teeth also was nearly if not quite 
completed, for in recent carnivores the lower carnassial 
is the first permanent cheek-tooth to appear. The best 
example at hand of a recent wild feline, showing the 
dental replacement, is a skull of the bay lynx (Lynx 
rufus), No, 0926, Y. P. M. Comparison of this lynx 
skull with that of Pogonodon serrulidens shows that, the 
replacement of the canines being at approximately the 
same stage in the two skulls, the replacement of the cheek- 

Eaton — John Day Felidce in Marsh Collection. 433 

teeth lias been relatively earlier in the fossil skull, for 
here the permanent cheek dentition is nearly or quite 
complete, while in the recent skull the upper deciduous 
carnassials and all the lower deciduous cheek-teeth are 
still in place and functional, although, on each side, the 
principal cusp of P 3 has just appeared, and the crowns 
of P 4 and M 1 have protruded to about half their height. 
If preferred, the cheek-teeth may be taken as the basis 
of comparison, in which case it would appear that the 
replacement of the canines was relatively tardier in the 
fossil skull than in the recent lynx. 

Cranium. — The basal length of the skull, measured 
between the basion and the probable position of the 
prosthion, is 167 mm. This shows the skull to be consider- 
ably shorter than that of Pogonodon davisi where the 
similar measurement is about 198 mm. Viewed from the 
side, the general proportions of the skull of P. serruli- 
dens resemble those of P. davisi more closely than they 
do those of Nimravus. This is especially noticeable in 
the low and straight nasal region, and in the frontal 
profile which rises higher from the base-line in the 
neighborhood of the bregma than in the region between 
the postorbital processes. The sagittal crest does not 
however, rise as high posteriorly in the present species 
as in P. davisi. 

The maxilla and the jugal, on each side of the skull, 
contribute to the formation of what may conveniently be 
termed the anterior zygomatic pedicle. It is believed 
that the widely varying proportions of this part of the 
skull will be found of considerable taxonomic value in the 
extinct Felidse. In certain genera the anterior zygomatic 
pedicle affords two excellent measurements, namely, 
a height, taken as the minimum distance between the 
inferior margin of the orbit and the alveolar margin 
of the maxilla, and a length, taken as the minimum 
distance between the external margin of the infra-orbital 
foramen and the posterior margin of the zygomatic 
process of the maxilla. By dividing the measured height 
by the length, and multiplying the quotient by 100, an 
index is obtained, admirably suited to the comparison of 
cranial form, since it is entirely independent of actual 
size. The anterior zygomatic pedicle of Pogonodon 
serrulidens is considerably higher than long, as it is in 

434. Eaton — John Day Felidce in Marsh Collection. 

P. davisi and P. platycopis, the indices of this part of the 
skull, in these three species, being 120, 148, and 162 
respectively, and so differing greatly from Nimravus 
where the length of the pedicle is equal to, or slightly 
greater than, the height, the corresponding index of 
N. gomphodus being 100, of N. dehilis 88, and of N. debilis 
major 92, while in Dinictis cy clops it is 106, and in J). 
squalidens (No. 8777, A. M. N. H.) it is 107. 

Fig. 3. 

Fig. 3. — Pogonodon serrulidens, sp. nov. Holotype. X %• 

Another remarkable characteristic of Pogonodon 
serrulidens is the large size of the infra-orbital foramen. 
The vertical and transverse diameters, taken within the 
foramen, not at its outlet, are respectively 13 mm. and 
10 mm. This is a variant toward the typical Hoplophonic 
form that at once distinguishes the present species from 
Nimravus. The true significance of very large infra- 
orbital foramina in certain groups of extinct Felidae is 
not clear. As the foramina are traversed by important 
branches of the external carotid artery and trigeminal 
nerve supplying the lower eyelid, side of the muzzle, 
and the upper lip, it may ultimately appear that the size 
of these foramina was directly correlated with the size 
and mobility of the upper lip in those extinct species that 

Eaton — John Bay Felichc in Marsh Collection. 435 

had long upper canines. It is not known to what extent 
the long knife-like upper canines of the true saber-tooth 
cats were curtained by pendulous upper lips when the 
mouth was closed in a state of rest : yet obviously such 
lips would have to be raised to avoid injuring them in 
the act of biting. Whatever the explanation, the fact 
remains that in the typical saber-tooth Smilodon the 
infra-orbital foramina are very large and in the false 
saber-tooth Nimravus comparatively small. 

The fragment carrying P 4 shows that the mandible at 
this section is much shallower and broader than the 
mandible of Nimravus. The vertical and transverse 
measurements of the mandible of Pogonodon serrulidens, 
taken immediately behind P 4 , are 23 mm. and 14 mm. 
respectively, and in P. cismontanus 32 mm. and 14.5 mm. 
respectively. I can not quote exact corresponding 
measurements in the types of the several species of 
Nimravus, but in other material, referred to that genus, 
the height is uniformly about three times as great as the 
transverse diameter. Cope evidently thought this 
character to be of taxonomic value, for he stated of 
Nimravus gomphodus 11 : "The ramus of the mandible 
is longer, deeper, and more compressed than in the recent 
species of Uncia and the Pogonodon platycopis" ; and 
of Pogonodon platcopis 12 he wrote "The mandibular 
rami are robust, and not so high and compressed as in 
Nimravus and its allies." Regarding the form of the 
anterior portion of the mandible in Pogonodon ser- 
rulidens, little can be ascertained from the fragment 
carrying the lower canine ; but the broad and flat anterior 
surface and the thickness of the bone external to the 
root of the canine are distinctly favorable to the sup- 
position that there was a flange for the protection of the 
upper canine. 

The foramina of the basicranial region, so far as it has 
been possible to locate them, occur as in Dinietis and 
Nimravus. The mastoid process, in point of size, 
resembles Nimravus rather than Dinietis, but it is 
directed a little more forward than in Nimravus, in this 
respect being more like Dinietis. The glenoid surfaces 

11 E. D. Cope, Report of the U. S. Geol. Survey of the Territories, vol. 3, 
965, 1884. 

12 E. D. Cope, Eeport of the U. S. Geol. Survey of the Territories, vol. 3, 
984, 1884. 

4:36 Eaton — John Day Felidce in Marsh Collection. 

are not projected downward as in Hoplophoneus, but, 
on the contrary, they rise considerably above the level 
of a line adjoining the basion and prosthion. 

The two fragments of limb-bones, already mentioned, 
have been compared with the corresponding parts of a 
puma, Felis concolor, No. 015, Y. P. M., whose skull is 
of almost exactly the same basal length as that of Pogono- 
don serrulidens. The lengths of the ulnae of the two 
animals can not be accurately compared, owing to the 
imperfection of the fossil bone, but a well-marked 
difference is presented in certain other proportions. 
While the shaft of the ulna of the fossil species is the 
more slender of the two, the diameter of its greater 
sigmoid cavity, measured in the axial direction of the 
limb, is considerably greater than in the recent species. 
This indicates also a greater diameter of the trochlea 
of the humerus. An analogous condition is presented 
by the fragmentary metatarsal III, the proximal 
articular surface of the bone being of almost exactly the 
same extent as in the example of Felis concolor, while 
the shaft of the fossil bone is slenderer. 

The present species is differentiated from Nimravus 
and Hoplophoneus by the aggregate of the cranial and 
dental characters, mentioned above, although, as might 
be expected, not by each of these characters severally. 
From Dinictis its differentiation is not so clearly denoted ; 
yet in consideration of the low nasal region and the 
relatively high parietal region, the reduction in size of 
upper premolars 2 and 3, the reduction of the inner root 
of P 4 with loss of protocone and addition of a well-marked 
parastyle, I cannot consistently assign to it a place with 
Dinictis felina, D. cyclops, D. squalidens, and D. pau- 
cidens. Pogonodon alone of known genera seems open 
to its reception, and the species to which, on the whole, 
it shows the closest affinity is Merriam's Pogonodon 
davisi. From this it is specifically distinguished by its 
smaller size, relatively lower sagittal crest, more com- 
pressed form of superior canines, 13 relatively greater 
diameter of the postorbital constriction, relatively lower 
and longer anterior zygomatic pedicle, and judging from 
Merriam's illustration of the teeth, by the lesser prom- 

13 The canine alveoli of P. davisi, although they do not furnish an exact 
basis for comparison, indicate a considerably greater transverse diameter. 

Eat oil — John Dag Felidce in Marsh Collection. 437 

inence of the inner root of the P 4 . Yet so close to 
Pogonodon davisi is the present species, that the two 
mnst share the same fate, should Pogonodon in the final 
analysis fail of recognition as a distinct genus or 

Dincelurus crassus, gen. et sp. no v. 

(Figs. 4-6.) 
Holotype, Cat. No. 10518, Y. P. M. Upper Oligocene (upper John Day), 
Turtle Cove, John Day Valley, Oregon. Collected by Wm. Davis in 1875. 

The type is a cranium without the mandible. It is 
well preserved, except that the posterior part of the 
sagittal crest and the contiguous parts of the lambdoid 
ridges have been destroyed and the postorbital process 
of the right frontal has suffered some abrasion. 

Fig. 4. 

Fig. 4. — Dincelurus crassus, gen. et sp. nov. Holotype. X %• 

Dentition. — I 3 , C 1 , P 2 , M 1 . The teeth are worn to a 
degree that indicates middle age, but not to such extent 
as to involve any modification of the alveolar parapet. 
The incisors, in their present condition, closely resemble 
those of Nimravns. Of the right canine only the base 
of the crown remains protruding from the alveolus. The 
left canine too has been broken, and a little of the base 
of the crown is lacking, but its characteristic form is 
shown sufficiently well to differentiate it from Nimravus, 
Dinictis, Pogonodon, and Hoploplioneus. Briefly charac- 

438 Eaton — John Day Felidce in Marsh Collection. 

terized, it is more truly feline than the canines of these 
genera, being more nearly elliptical in section through- 
out its length and less flattened on the inner surface. 
The maximum (antero-posterior) diameter of the canine 
at the alveolar margin is 15.5 mm. and the transverse 
diameter 11 mm. Dividing the transverse diameter by 
the maximum diameter, and multiplying the quotient by 
100, an index of 71 is obtained. The corresponding 
indices of the superior canines of several well-known 
species of extinct Felidse are as follows : Nimravus 
gomphodus 47, derived from Cope's measurements; N. 
debilis 57 ; N. debilis major 56 ; Pogonodon platycopis 
54 ; P. davisi 52, derived from alveolar diameters ; 
Dinictis cy clops 64; D. felina 47. The diameters of the 
superior canine of Pseudcdurus quadridentatus , at the 
level of the neck, given by Filhol, 14 transverse 8 mm. and 
antero-posterior 12.7 mm., yield an index of 71. In 
examples of three recent feline species I find the corre- 
sponding indices to be: Felis tigris 73; Felis concolor 81; 
Cyncelurus jubatus 79. The vertical length of the canine 
in the present species cannot be exactly determined 
because of its imperfection. Its length, however, 
appears to have been about the same, relative to the 
length of the skull, as in Nimravus debilis; and relatively 
less than in N. gomphodus. Except in young adult 
animals the length of the canines is of course an unsatis- 
factory quantity. The anterior ridge of the tooth lies 
further back from the actual anterior margin than in 
Nimravus, and in this respect also the present species is 
further advanced and more cat-like. The posterior ridge 
is not continued along the neck of the tooth to the alveo- 
lar margin. This has much to do with the broad, ellipti- 
cal section of the neck of the tooth. There is no trace, on 
either side, of P 2 or its alveolus. For this reason two 
premolars only are postulated in the dental formula 
stated above. If in early adult life there was a P 2 , it can 
hardly have been other than exceedingly small and 
vestigial, as the shortness of the postcanine diastema, 
9.5 mm. on one side and 10 mm. on the other, renders the 
presence of a P 2 of any considerable size highly improb- 
able. The shortening of the postcanine diastema with 
reduction of P 2 to vestigial size would have practically 
the same significance as the shortening of the diastema 

14 H. Filhol, Annales des Sciences Geologiques, 21, 76, 1891. 

Eaton — John Day F el idee in Marsh Collect ion. 439 

with total loss of P 2 ; either condition would denote 
a variation from the more complete dentition of 

P 3 is a large tooth, its principal cusp being sub-equal 
in size to the paracone of P 4 , as in Nimravus debilis. 
From the last named species it differs in having a slightly 
greater transverse diameter of the crown relative to the. 
antero-posterior diameter, and the inner division of the 

Fig. 5. — DincElurus crassus, gen. et sp. nov. Holotype. X %• 

posterior root is much stouter, a condition that suggests 
further advance toward Pseudcelurus and Fells. P 4 
exhibits less development of the inner root than I have 
observed in any specimens of Nimravus, and the inner 
root does not support a positive protocone. There is 
less trace of a parastyle than in Nimravus, and the 
external outline of the crown is slightly convex instead 
of being slightly concave as is usually the case in 
Nimravus. This convexity of crown gives the tooth an 

appearance of 


breadth, if the buttress formed 

44:0 Eaton — John Day Felidce in Marsh Collection: 

by the inner root be disregarded, than in Nimravus. M 1 
is a very small tooth, and the reduction of its inner lobe 
has proceeded so far that the transverse diameter of the 
tooth-crown is not greater than the antero-posterior 
diameter. Since the external margins 'of the premaxillse, 
between the canines and the lateral incisors, are widely 
and deeply concave, when viewed from above or below, 
it may be inferred that the lower canines, which were 
given clearance by these concave diastemata when the 
jaws are closed, were also of larger transverse diameter 
than those of Nimravus; and this inference is supported 
by fragments of a lower canine that were found embedded 
in the matrix. So deeply concave are these precanine 
diastemata that they extend further to the rear than 
the anterior margins of the maxillary parapets enclosing 
the canine alveoli. 

Cranium. — The skull in its general proportions shows 
a greater resemblance to Nimravus and Dinictis than to 
Hoplophoneus. From Pogonodon, which is in some 
respects intermediate between Dinictis and ■ Hoplo- 
phoneus, it is not so distinctly separated by cranial pro- 
portions as by dentition. The foramina of the 
basi-cranial region, the height of the glenoid articulations 
relative to the base-line of the skull, and the form and 
position of the mastoid processes are essentially the same 
in the present species as in Nimravus; but in the wide 
spread of its zygomatic arches and in the breadth and 
massiveness of the entire fore part of the skull — charac- 
ters that have suggested the specific name — the present 
species departs widely from the proportions of the 
species of Nimravus and Dinictis. 

With a basal length of 174 mm., measured from basion 
to prosthion, the bizygomatic diameter is 164 mm. The 
relation of these measurements is expressed by the bizy- 
gomatic index of 94, which is well outside the range of the 
corresponding indices of Nimravus, Dinictis, and Pogono- 
don. The bizygomatic indices of the best known species 
of the three above-named genera are as follows : Nimravus 
gomphodus 72; N. debilis 76; N. debilis major 71; 
Pogonodon davisi 74 (approx.) ; P. platycopis 78 
(approx.) ; P. serrulidens 11 (approx.) ; Dinictis cy clops 
85 ; D. squalidens, No. 8777, A. M. N. H, 82 ; D. felina 
76; D. paucidens 79; D. fortis 82 (from illustration). 

Eaton — John Day Felichc in Marsh Collection. 441 

Similarly the interorbital breadth of the present species 
exceeds the corresponding measurements in Nimravus, 

Dinictis, and Pogonodon, as may be seen from the com- 
parison of the interorbital indices, computed from the 
interorbital diameter and the basal length. The inter- 
orbital breadth index of Dincelurus crassus is 35 while 
that of the other well-known speices is as follows : Nimra- 
vus gomphodus 25 (approx.) ;. N. debilis 26; N. d chilis 

Fig. 6. 

Fig. 6. — Dincslurus crassus, gen. et sp. nov. Holotype. X %• 

major 24; Pogonodon davisi 28 (approx.) ; P. platycopis 
27 (approx.) ; P. semdidens 28 (approx.) ; Dinictis 
cyclops 33; D. sqiialidens, No. 8777, A. M. N. EL, 28; 
D. fortis 33 (from illustration). 

The diameter of the postorbital processes of the frontal, 
which here gives an index of 46 when contrasted with the 
basal length, is relatively mnch greater than in Nimravus 
where the range is from 34 to 36. In Pogonodon this 
index has a wider range, as the following values show: 

Am. Jour. Sci — Fifth Series, Vcl. IV, No. 24. — December, 19C2. 

44:2 Eaton — John Bay Felidce in Marsh Collection. 

CO lO H^ CO i— I rH 

•aoti *ds ^a 'iraS 'snssvjo snmi30u%(j r " 1 r ~ l ,H 



•ds l su9p%\niids uopouodoj HHH 

H< ?D H< oo TtH co oq 

lO lO - kO CD lO 

a c3 UJ . " J . ce "■•>. ^ u ? rf ^ 

7TO -atoII '68Z 'ON f 8d^o T o H g g £ iS^aSSwNSqS 

•Ji8pj ismvp uopouoffoj ^ <*> ^ g 

*H *N *H -Y '8869 *0N '^oton lll^lggefgio^o^ 5 

•(odoQ) stdooftivid uopouodoj <^> ^ ,H ^ >^ 

c3 n3 

.•9d^ p S S^pprj I a I JHCO «|jg«> 3 

•A*pi8rj_ VUipJ S%10%WI(X rH rH r-l g 


•H ".*[ -W.-V 'ZZZ8 'ON |g£ £ I £ II * 

•(ado£)) su9p%\vnbs si%o%u%(j «h th i-t © 

•H 'N 'H 'V 'Z869 "ON '9<K*oioh S^Sol^oggo^ololS 00 g 

•edoQ sdopfio si^oiuiq; <h .h rn ^ 


'TAT ',T " X 'QTzOflT 'OM t^iocoot-HHCJr-fooooascoHHao 

jal a a yyuu l "iv tot-oqcqH^cocooooqHHoqcorH r-T 

•ds snavj,ULij$ n th th th 

lO rH 

"TAT */T 'X '/TQ0T "0\T oooow^^SHiroaH^oooao © 

•(adoQ) s%\%qpp snaviuiift th oq th ,d 

pq ^ '(adoQ) siiiqsp sncwjLWijx oq oq th g 

M ° c3 * ^ *° -«J 

En § '(edoo) sipqop snavjbwifl; rH oq th jg 

"JTTO * A T n £L '1891 '0.K 'sdA^opH H >° 

Ol lO 

•(•ii8]/\[) dofvw siiiqgp snchv^wijy; oq oq rH 

'H "NT *W 'V '0869 'OKT ^d^iOTOH coocoOicocot-HHOoocsicoH^oo 


•(8do[)) s%\%qpp sniMUwijx rn rH rn ^ 

'H 'K 'W V '8869 'ON 'ad^ojpH c»©|oopoo 

•sdoQ snpoi[duio6 snciviwi^ 



ft fl © s s § 

rg-g ^^ gr.,5 


„ m ri M _i 

+3-^2 r-i 2 bJO-3 cS c3 cS or fl r =5 "S 

§■&■■§ 's^^gg | ^ch«h § I .a s 

•s^uaiue.ins^aTn jo -son pwag 

rx .2 >■> S? © S -jf "oS ^j +?' 'oJD &C . . ft m 

rew.,_ijH ri . rt oajno L ©orJ3 -■ © 


ft 3 


Eaton — John Bay Felidce in Marsh Collection. 443 

*H 'J 'A '8TS0T 'or 'ad^opH ^ 

•aoh 'ds q.a -uaS ' siunixuid °* 

•J\l *d "A 'OoSOI '°iM 'ad^oiOH b- 

•aou *ds 'suQpiprujLOs uopouolwj ^ 

711^0 ' A T n £L '6SZ 'on: 'sd^oioH th 
M.i8j\[ isiavp uopouoOoj; 

'H '.NT 'H *Y '8869 *0X 'adA>o t oH 
•(edoQ) sidoofiivid uopouoGoj 

•adX; pg s^pxaq; 







i— 1 

















*H 'W "H 'V 2ZZ8 'ox cm 
(adoQ) suQpypmbs siwiuiq; °° 

•H '.N "H 'V '2869 'OK 'sdAoioH 10 

•adoQ sdopfio si^diuiq; °° 

*H *<J 'A '9KI0T '0X t~ 

•ds sncivjbwifi t - 

"H *d 'A 'ZTSOI 'OX <m 

•(gdo^) sii%qpp snavjwift *>■ 

*H 'd *A '2^001 'ON eq 

•(adoQ) s%i%qpp sncivu,m,ij^[ ^ 

•R 'J 'A '^001 'ON w 

•(adoQ) s%i%q9p sncivjamjx t- 

7TTO ' A T lI fl C T89I *ox 'edAoioH ^ 

(\u:8j\[) jofvui sfijqjp snavjiwifii ^ 

'H "X 'K -V '0869 "OX 'adA>i<>H 

•(adoQ) s%i%q?p snavdULijg o- 

"H "M *H 'Y '8869 *on 'adA>i<>H N 

•adoQ snpoi[dwo6 snavduiifl b- 







































r- 1 






























« CO 





■?. ?U 










X oc 





O ! 






O I 






. , rH 







<* | 






-t- 1 
















, | 




• Pi 




























444 Eaton — John Day Felidce in Marsh Collection. 

Pogonodon davisi 43 (approx.) ; P. platycopis 31 
(approx.) ; P. serrulidens 36 (approx.). From the follow- 
ing values of this index for Dinictis, it will be seen that 
only one species of this genus is known to have a 
relatively greater postorbital breadth than Dincelurus 
crassus: Dinictis cy clops 51; D. squalidens, No. 8777, 
A. M. N. H., 35; D. felina 34; D. fortis 37 (from 

The breadth of the anterior nares, relative to the basal 
length, when reduced to an index of 21, is also greater 
than in the other species with which it has been compared 
in Table B. Pogonodon platycopis with an anterior 
narial index of 19 approaches the present species most 
closely in this character, but this index has been computed 
for only a few type skulls. The breadth of the postorbi- 
tal constriction, which may in like manner be expressed 
by the index 25, separates the present species from 
Nimravus, but not from all the species of Dinictis and 
Pogonodon, Dinictis cyclops having a corresponding 
index of 27 and Pogonodon serrulidens of 25 (approx.), 
while the highest value in Nimravus is an index of 21 
for N. debilis. The bimastoid diameter is relatively 
greater than in Nimravus gomphodus, N. debilis, 
Pogonodon platycopis, P. serrulidens, and Dinictis squali- 
dens (No. 8777, A. M. N. H.), but slightly less than in 
Dinictis cyclops and D. felina. With regard to the 
bicondylar diameter, the series of type skulls, on which 
this measurement has been taken, is too limited to admit 
of any definite statement about the generic value of the 
bicondylar index. 

In conformity with the remarkable breadth of the fore 
part of the skull, the palate also is broader than in the 
allied genera with which comparison has been made. 
This is especially noticeable in the posterior palatal 
breadth as expressed by the transverse measurement 
between the inner roots of the upper carnassials. In 
the present species the index derived from this measure- 
ment and the palatal length is 72, which greatly exceeds 
the corresponding indices computed for Nimravus debilis, 
N. debilis major, Pogonodon davisi, P. platycopis, P. 
serrulidens, and Dinictis cyclops. The species that 
approaches nearest in this respect is Nimravus debilis 
major with an index of 59. The three species of Pogono- 
don and Dinictis cyclops fall short of this last value. 

Eaton — John Day Felidce in Marsh Collection. 445 

For the other species of Dinictis this index has not been 
computed, but, judging from published illustrations none 
of them appear materially to exceed D. cyclops in the rela- 
tive posterior palatal breadth. The entire posterior por- 
tions of the maxillae, including the zygomatic processes, 
are exceedingly massive, and the palatine processes of the 
maxillae, instead of forming thin, sharp-edged orbital 
floors, as in the examples of Nimravus, Pogonodon, and 
Dinictis that I have examined, are thick and rounded 
posteriorly. The palatal vault is much deeper than in 
Nimravus. This is caused, not by a more pronounced 
arching of the palatines and the palatine processes of 
the maxillae, but entirely by the greater vertical depth 
of the alveolar processes which form the sides of the 
vault. A direct result of this modification is seen in the 
almost complete elimination of the circular pits, in the 
palatine processes of the maxillae, that in Dinictis, Nimra- 
vus, Pogonodon, Pseudcelurus, and Hoplophoneus receive 
the principal cusps (protoconids) of the lower carnas- 
sials when the jaws are closed. Slight irregularities in 
the contours of the maxillae, at the lowest points of origin 
of the masseter muscles, are by no means uncommon in 
the Felidae, recent and extinct, and osteological peculiari- 
ties of this description would probably be most noticeable 
in middle-aged individuals of large size and robust form. 
Accordingly caution must be exercised lest too much 
importance be given to differences that may be due solely 
or largely to individual variation. It would, however, 
be a serious omission to fail to note the form of the 
maxillae immediately external to the superior molars in 
the present species. On each side of the skull, the power- 
ful development of the masseter muscles has apparently 
caused changes in the maxillary contours amounting 
almost to the production of bony masseteric processes. 
The lack of symmetry, displayed on the two sides, 
must of course be attributed to individual variation, yet 
so pronounced is the smaller outgrowth on the left side, 
that, in view of the unusually massive character of the 
entire fore part of the skull, it may properly be regarded 
as of at least specific value. In this connection, the 
following quotation from "Wortman's 15 description of 
Dromocyon vorax is of interest: "On the under surface 

15 J. L. Wortman, This Journal (4), 12, 293, 1901. 

446 Eaton — John Day Felidce in Marsh Collection. 

of the arch, at the point of junction of the molar with 
the maxillary, there is a prominent process for the 
tendinous origin of the most anterior fibres of the 
masseter. This process is apparently wanting in the 
carnassident skull, but is large and prominent in that of 
the Opossum, Dasyure, and to a less extent in 

One of the most distinctive characters of this skull 
is the form of the portion of the maxilla that, for want 
of a better name, I have termed the anterior zygomatic 
pedicle. The minimum height from the alveolar margin 
to the inferior margin of the orbit is 46 mm. and the 
minimum length, measured on the outer face of the 
maxilla, from the infra-orbital foramen to the posterior 
margin of the zygomatic process, is 30.5 mm. Dividing 
the first of these measurements by the second, and multi- 
plying the quotient by 100, an index of 151 is obtained. 
Inspection of Table B for allied genera and species will 
show that by the proportions of this part of the skull, 
the present species is easily distinguished from Nimravus 
and from Dinictis cyclops, D. squalidens , and D. felina. 
Judging from Riggs' figure of the skull of Dinictis 
paucidens 16 and Adams' figure of D. fortis, 17 the anterior 
zygomatic pedicle in these species also is low and long, 
as in D. felina. 

The glenoid fossae are projected no further below the 
basicranial plane than in Nimravus. The mastoid and 
paroccipital processes are similar in size and direction 
to those of Nimravus. The basicranial foramina occur 
as in Dinictis and Nimravus, with this important 
exception that in Dincelurus crassus the posterior opening 
of the carotid canal is a veritable foramen, completely 
enclosed by the thin lateral expansion of the basi-occipi- 
tal, and not merely an open notch in this lateral expan- 
sion, as is the case in various examples of Nimravus, and 
as represented in Cope's figure of the type of Nimravus 
debilis. 18 Scott's illustration of the skull of Dinictis 
felina 19 might be supposed to indicate that the carotid 

16 Elmer S. Biggs, Kan. Univ. Quar., 4, No. 4, 239, 1896. 
"'Geo. I, Adams, Am. Nat., 29, pi. XXVI, 1895. 

18 E. D. Cope, Eeport of the U. S. Geol. Survey of the Territories, vol. 3, 
952, 1884. 

19 W. B. Scott, Proc. Acad. Nat. Sci., Philadelphia, 41, 213, 1889. 

Eaton — John Day Felidce in Marsh Collection. 447 

canal passes through the basioccipital, but I am inclined 
to believe that such is not the regular occurrence in 
Dinictis, for in Matthew's detailed figure of the basi- 
cranial region of Dinictis, 20 presumably drawn from the 
excellent specimen of Dinictis squalidens , No. 8777, 
A. M. N. H., the entrance to the carotid canal is appar- 
ently through an open notch in the lateral expansion of 
the basioccijntal, and precisely the same condition has 
been observed in a skull in the Marsh Collection, Yale 
University, Cat. No. 10048, identified as Dinictis squali- 
dens. The otic bullae, although incomplete, are much 
more fully developed than in Nimravus, a considerable 
portion of the thin wall of the bulla, on each side, being 
preserved in extension of the thickened tympanic ring, 
and traces of the median walls of the bullae being visible, 
closely pressed against the lateral expansions of the 
basi-occipital. In this respect the present species shows 
a decided advance beyond the stage reached by Nimra- 
vus; and a like further development toward the recent 
Felidae may be observed in the shortened length of the 
alisphenoid canal and in the slenderer proportions of the 
bridge of the alisphenoid bone, protecting the external 
carotid artery at this point. The brain-chamber appears 
to have been a little larger, in proportion to the basal 
length of the skull, than in Nimravus, but this superiority 
may be more apparent than real, since external measure- 
ments only are available, and it is possible that the cran- 
ial walls of the present species, in keeping with the 
robust development of the skull generally, may be slightly 
thicker than in the species with which comparison has 
been made. 

The fossil skull and the hard ashy matrix are very 
light gray in color, and so appear to satisfy the require- 
ments of the upper John Day beds, as they occur in the 
Cove on the John Day River, where the record states 
the specimen to have been found. This reference of the 
type to the upper John Day is further supported by the 
advance, shown in certain cranial and dental characters, 
beyond the stage of evolution denoted by Nimravus 
debilis and Nimravus debilis major, both of which species 
are understood to be from the middle John Day. 

20 W. D. Matthew, Bull. Am. Mus. Nat. Hist., vol. 28, 299, 1910. 

448 Eaton — John Day Felidce in Marsh Collection. 

Genus Nimravus. ' 

Four of the crania in the Marsh Collection, whose 
measurements and indices have been included in the 
tables, require but brief mention. Those listed under 
the numbers 10044, 10045, and 10517 have been referred 
to Nimravus debilis because on the whole, although they 
differ mutually in minor characters, they resemble A T . 
debilis and N. debilis major rather than N. gomphodus. 
It is true that these three crania have almost exactly the 

Fig. 7. 

Fig. 7.— Nimravus debilis (Cope). Cat. No. 10517, Y. P. M. X %■ 

same basal length as the type of the latter species, viz., 
198 mm. ; but the argument, founded on size alone, fails, 
since within the species N. debilis, the basal length has 
a range variation from 176 mm. to 211.4 mm., as shown 
in Table A. 

Three views of the cranium designated as No. 10517 
are given in figures 7, 8, and 9, and two views of the 
cranium No. 10045 appear in figures 10 and 11. The 
mandible of the latter is the better preserved of the two, 
and is believed to show the form of masseteric fossa 
regarded by Merriam as characteristic of N. debilis and 
N. debilis major. The following is quoted from 
Merriam 's painstaking discussion of Cope's Nimravus 
gomphodus and N. (Archcelurus) debilis : "Practically 
the only character which seems distinctive is found in 

Eaton — John Day Felidce in Marsh Collection. 449 

the form and size of the masseteric fossa. In N. gompho- 
dus its inferior margin is separated from the lower 
border of the horizontal ramus by a wide bar. In A. 
debilis the fossa extends farther forward and reaches 
down to the inferior margin of the jaw, which it may 
follow for some distance." 21 Cranium No. 10045 is in 


Fig. S.—Nimravus debilis (Cope). Cat. No. 10517, Y. P. M. X %• 

many respects better preserved than No. 10517, although 
the condyles have been lost, and it shows admirably the 
positions of the basicranial and palatal foramina, and 
also the delicate tympanic ring of the right side. 

As regards the size of the infracarnassial. exostoses, 
it may be said that in the mandibles of the three individ- 
uals, Nos. 10044, 10045, and 10517, these outgrowths are 
strongly developed. No. 10045 originally possessed four 
superior premolars on each side, P 1 on the right side 
being double-rooted. In the mandible of this specimen 
the first and second premolars have been lost, the only 

21 J. C. Merriam, op. cit., p. 43. 

450 Eaton — John Day Felidce in Marsh Collection. 

Fig. 9. 

Fig. 9. — Nimravus debilis (Cope). Cat. No. 10517, Y. P. M. X %• 

Fig. 10. 

Fig. 10.— Nimravus debilis (Cope). Cat. No. 10045, Y. P. M. X % 

Eaton — John Day Felidce in Marsh Collection. 451 

trace of any of these teeth being the nearly closed 
alveolus of P 2 on the right side. Jnst how much reliance 
can be placed on the diameters of the superior canines, 
as specific characters, is uncertain ; still, attention should 
be called here to the fact that, in the diametral indices of 
the superior canines, the three specimens, Nos. 10044, 
10045, and 10517, agree much more closely with the types 

Fig. 11. 

Fig. 11.— Nimravus debilis (Cope). Cat. No. 10045, Y. P. M. X % 

of N. debilis and N. debilis major than with N. gompho- 
dus; and the same may be said with reference to the 
height index of the anterior zygomatic pedicle. 

The fourth skull, No. 10046, Y. P. M. (fLg. 12), whose 
measurements and indices are given under the genus 
Nimravus, is much smaller than the threa last enumer- 
ated, being in fact even smaller than the type of N. debilis. 
The superior canines are relatively longer than in that 
type, yet their measured diameters are almost exactly 
equal to those of N. debilis. The postcanine diastemata 
are short; and in the reduction of the premolar teeth, 
as well as in the slight development of the infracar- 
nassial exostoses, this specimen differs from the type of 

452 Eaton — John Day FelidcB in Marsh Collection. 

N. debilis. Viewed in profile, the anterior surface of the 
mandible appears to rise more nearly at a right angle 
with the horizontal ramus than in the types of N. debilis 
and N. gomphodus; and this squarely truncated appear- 
ance of the mandible is increased by the development of 
a prominent mental rugosity. Labels accompanying 
this skull show that, at one time, it was provisionally 
identified as N. gomphodus, at another time as N. debilis. 
On the whole its indices point to a closer affinity with 
N. debilis than with N. gomphodus. The skull of Cope's 

Fig. 12. 

Y. P. M. 

N. confertus is but little smaller than No. 10046 ; but as 
N. confertus is represented only by a very imperfect 
mandible, a satisfactory comparison of the specimens 
is impossible. There is, however, a remarkable simi- 
larity in the form of their mandibular symphyses. In 
each the symphysis is very short horizontally and extends 
but little to the rear of the posterior margins of the lower 
canine teeth. No. 10046 is supposed to have been found 
in the middle John Day. 

Hubbard — Antimony Mines of Shiu Chow. 453 

Akt. XXXVII. — The Antimony Mines of Shiu Chow, 
China; by Geoege D. Hubbaed. 

Introdiictiou.^Slim Chow is a walled city in the north- 
ern lobe of Kwangtung province, where the boundary 
between the two provinces of Hunan and Kiangsi comes 
south to their more famous neighbor, Kwangtung. It 
is about 30 miles from the Kwangtung border both west, 
northwest, and north, where the latter curves around 
in the midst of rather mature but little used mountains. 
It stands at a fork in the Pei Kiang, or North River, 
which has opened up a valley leading south to Canton 
and Hongkong. A railroad has also been constructed 
from Canton northward as far as Shiu Chow in its reach 
to connect with Changsha and Hankow. 

Topography. — : Shiu Chow has grown up in topography 
a little more mature than that along the provincial border 
west and north, but not nearly so old as that which has 
been partly submerged to make the great fingered and 
island-spattered Canton bay, into which both Pei Kiang 
and Si Kiang or West River flow. 

Stratigraphy. — The rocks of the region are very deeply 
weathered Paleozoic limestones, sandstones, and shales, 
probably ranging in age from Ordovician to Carbon- 
iferous inclusive. The strata have been elevated in rather 
closely pressed folds whose trend is practically north 
and south, or perhaps north-northeast and south-south- 
west. While in many places within a few miles east and 
south the rocks dip more gently, all about the mines they 
dip very steeply, usually over 80°. Strikes observed at 
several points range from N 10° E to N 20° E. One can 
get strikes of very diverse angles in this region, but those 
departing far from the above are local. The trend of the 
big limestone ridges very nearly north and south is quite 

Beginning some 5 or 6 miles north of Shiu Chow and 
just west of the mines (iig. 1), the rocks crossed in a sec- 
tion eastward for a mile or more do not seem to repeat at 

Provisionally, the massive, dark blue, calcite-veined 
limestone, west of or above the mines, has been called the 
Ordovician. The evidence is both in its position with 

454 Hubbard — Antimony Mines of Shiu Chow. 

reference to the succeeding beds to the east, including the 
coal, and in its fossil content. A few fossils of gastro- 
pods and brachiopods have a distinctly Ordovician f acies, 

Fig. 1. 

Fig. 1. — Geologic map, Shiu Chow to antimony mines. 

though no species are identified. I saw none of the big 
cephalopods so commonly found in the Ordovician in 
other parts of China, but learn from Chinese observers 

Hubbard — Antimony Mines of Shin Chow. ±00 

that they occur near here. (Exact location and horizons 
unknown.) This big limestone makes a rather rugged 
ridge of very uneven height, yet continuous southward to 
the west branch of the river and several miles beyond. 

A shale bed. probably as thick as the limestone, follows 
the latter and. is succeeded by a qnartzitic limestone. The 
shales give low. open country, but the quartzite makes 
a small ridge. Xext comes a thick series of shales of 
many colors. — blues, greens, reds, and browns. Some 
beds of this series are a little more resistant than others, 
but none of them are strong enough to hold up ridges. 

Farther east than the shales and about a mile from the 
blue limestone is another ridge-maker, parallel with the 
first. Its beds are thinner and mostly gray, not blue, 
and not as dark as the former, but violently contorted 
and plicated. It has a few veins of calcite, but no such 
development as the other limestone has. Xo fossils 
were seen in any of these series except in the western 
limestone, but no search could be made because of lack of 

Structure. — Coal is reported a short distance east of 
this second limestone and is mined 6 or 7 miles down- 
stream, which would probably be not more than 5000 
feet east, stratigraphically, of the contorted limestone. 
In the vicinity of the coal, the dips in shales and sand- 
stones are S. E. by E. and not strong, while in the lime- 
stones and quartzites above Shiu Chow, nearly everything 
is about on end. Time did not permit scouting to the 
west to ascertain certainly, but the structure has the 
appearance of a great anticline, the east half of which is 
two miles or more across. The rocks in the west part 
of the sections would then be the oldest seen, and the coal 
and associated rocks the youngest. The most probable 
interpretation for the ages of the rocks, then, is that the 
western or dark blue is, as suggested, Ordovician; the 
lower shales may be Silurian; the quartzites and upper 
shales then would follow as Devonian; and the contorted 
limestone may be Mississippian, with Coal Measures 
above or on eastward. 

Ores. — For several years, in the stream beds leading 
down eastward from the older limestones, the peasants 
have been picking up pieces of stibnite and more or less 
oxidized mas>es of antimonv ore. In the summer of 1920, 

456 Hubbard — Antimony Mines of Shiu Chow. 

Mr. S. P. Chen became interested in this locality and 
began digging to find the ore in place. Several prospect 
pits and a shaft 25 feet deep were made. Ore was f onnd 
in most of them in float, and still more was fonnd in the 
stream beds, especially after rainy seasons. Finally, 
when prospecting had shown the more favorable course 
to pursue, a horizontal tunnel was directed straight into 
the hill from a point well up, but believed to be safely 
below the source of the ores. Before this tunnel was 
extended 100 feet, it struck the stibnite body in the lime- 
stone. The ore body is at least 3 or 4 feet thick, seems 
to be nearly perpendicular, though it probably dips with 
the rocks, hence 80° to 85° eastward. 

The distribution of the float ore indicates that the vein 
of ore runs more or less continuously along the top of the 
limestone beds and below the shales. This contact seems 
especially favorable for ore deposition. The author sug- 
gests that the deposits may have been put in place as a 
vein before the folding occurred, or in part a replacement 
accumulation in the limestone, and that the less pervious 
beds of the shales above may have prevented the waters 
from rising higher and thus localized the deposition of 
the ore. No igneous rock could be found in the vicinity, 
and inquiry of the operator and others brought the uni- 
versal testimony, "No granite; all limestone, shale and 
sandstone." In the impregnated limestone occur occa- 
sional little clusters of tiny pyrite crystals. 

Toward the top, the vein is weathered, and the 
weathered products of stibnite occur. The only primary 
ore is this sulphide of antimony; the weathered ores are 
rarely stained more than a tint, confirming the belief in 
the absence of pyrite. 

Some 12 to 13 miles south to southwest of these works, 
the float ore has also been found and followed up by the 
same exploiter, but so far, no primary deposits have been 
found. These finds are along the same limestone beds, 
but 7 or 8 miles south of Shiu Chow. 

In its intimate associations, the ore varies from almost 
pure stibnite to stibnite and calcite gangue with only 
small per cents of the sulphide. The mixed ore is very 
pretty, for the calcite is coarse-grained, crystalline, and 
nearly clear, and is thrust through in every direction by 
the prismatic crystals of shiny metallic stibnite. The 
best stibnite ore runs about 65% antimony. Pure stib- 

Hubbard — Antimony Mines of Shin Choiu. 457 

nite, Sb 2 S 3 , carries 71.4% of metal. The black, less 
crystalline ore runs about 50%, but those rich in the 
gangue calcite can be worked down to 30% antimony. 
The oxide ores, mostly stibiconite, produce 50 to 55% 
antimony, not a very pure ore. 

About 33 miles north b}^ northwest from Shiu Chow 
are small native workings for antimony ore in float and 
in stream beds. Some of this is practically pure stibnite. 
It is brought to Shiu Chow to be smelted. 

Mining and Transportation. — All the work in this 
whole region is carried on by the simple, laborious, 
native methods. Ore and dirt are moved in baskets on 
carrying poles or carried